Development of computational tools for designing water and energy nexus: Novel applications of multidisciplinary optimization and machine intelligence

投资于高度分散的可持续能源的大规模开发，可以同时解决我们的能源问题，我们的气候问题（就减少我们对气候变化的影响而言），同时为大量的人提供就业机会，帮助改善经济，提高公平。今天，可持续（可再生）能源主要来自水电、风能、太阳能和生物燃料。这些资源依赖于水和土地。过去用于设计这些系统的大多数工具都没有考虑到全球气候（和变化）、水文和当地地理（例如安大略省的大小）、能源生产的不确定性以及当地/全球的反馈。这些目标要求发展以数学模型和数值解决方案为支持的方法，使用气象、水文、地理和社会需求的大数据，这些都属于水-能源关系的背景。有效地解决这些问题需要集成建模工具，例如计算效率的代理优化，这可能涉及基于机器智能的昂贵仿真模型的函数逼近，模拟各种人类活动的基于代理的建模，以及基于梯度和无梯度的优化算法。申请人在这些方法方面的专业知识将被结合起来，帮助解决现在和未来的可持续能源问题。可持续发展需要长期的全球视野。此外，实际的社会需求要求考虑当地的短期成本和收益。安大略省是加拿大人口和面积都很大的一个省，它的非碳能源生产严重依赖核能（占总发电量的60%以上）。核能的未来尚不清楚，然而，安大略省在可再生能源方面也有巨大的潜力，该提案旨在填补大规模实施可再生能源所需的建模和设计方法方面的空白，包括抽水蓄能水电系统（比电池更便宜的储能替代品，但仅在特定地点可用），以增强可再生能源的间歇性发电。可再生能源对环境的依赖程度高，但不可再生能源对环境的影响很大。需要专门的模型和方法来考虑水和能量的扩散和集中性质。我们将发展能够同时考虑到以可持续性为目标的能源和水的人类使用的方法，为这种发展规划出巨大的潜在领域。方法可推广到其他省份的可再生能源发展，也适用于其他大规模的环境问题。

Investing in a large scale development of highly distributed sustainable energy can simultaneously solve our energy problem, our climate problem (in terms of reducing our impacts on changing climate) while giving employment to a large number of people that help improve the economy while improving equity. Today, sustainable (renewable) energy comes mainly from hydropower, wind and solar energy, and biofuels. These resources depend on water and land. Most tools that were developed in the past for designing these systems didn't take into consideration global climate (and changes), hydrology and local geography (e.g. the size of Ontario), uncertainty in energy production, and local/global feedbacks. Such objectives call for the development of methods supported by mathematical models and numerical solutions using Big data on meteorology, hydrology, geography, and societal requirements and these fall in the context of the water--energy nexus. Solving these problems efficiently requires integrated modeling tools together such as surrogate optimization for computational efficiency, which may involve machine intelligence -based function approximation of costly simulation models, agent-based modelling for simulating diverse human activities, and both gradient--based and gradient--free optimization algorithms. The applicant's expertise in these methods will be combined to help solve sustainable energy problem for now and the future. Sustainability demands long-term global outlook. Additionally, practical social requirements demand consideration of local short-term costs and benefits. Ontario, a large province both in population and size in Canada is heavily dependent (over 60% of the total energy generated) on nuclear energy for its noncarbon based energy generation. The future of nuclear is unclear, however, Ontario also has huge potential for renewable energy and this proposal aims to fill the gap that exist in modelling and design methodologies required for large scale implementation of renewable energy including pumped storage hydropower systems (cheaper alternatives to batteries for storage but available only in specific locations) to enhance the intermittent generation of renewable energy. Renewable energy is highly dependent on environment as its source, but non-renewable energy is highly impactful on the environment. Specialized models and methods are necessary to consider both the diffusive and the concentrated nature of water and energy. We will develop methods that can simultaneously consider human uses of energy and water guided by sustainability as the goal to map out large potential areas for such development. Methods are extendible to other provinces for renewable energy development but also for other large scale problems in the environment.