Uncertainty quantification for adaptive surrogate modeling framework using CFD simulations****

使用 CFD 模拟对自适应代理建模框架进行不确定性量化****

• 批准号: 537813-2018
• 负责人: Ponnambalam, Kumaraswamy
• 金额: $ 1.82万
• 依托单位: University of Waterloo
• 依托单位国家: 加拿大
• 项目类别: Engage Grants Program
• 财政年份: 2018
• 资助国家: 加拿大
• 起止时间: 2018-01-01 至 2019-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=662425
• 关键词: Uncertainty; quantification; adaptive; surrogate; modeling

The goal of this project is to develop a probabilistic framework for uncertainty quantification and propagation for surrogate models generated to emulate the expensive computational fluid dynamics (CFD) solver incorporated in a physics-based life cycle management expert system (LCM-ES) framework called XactLIFE (TM) developed in-house by the industrial partner (LPTi- Life Prediction Technologies Inc.). LPTi provides prognostics based life cycle management services for gas turbine engine parts utilizing the XactLIFE system. As part of a new product offering to their clients, LPTi aims to custom build the XactLIFE system for specific engines with preselected critical parts. This would need the substitution of the CFD solver present in the XactLIFE system with surrogate models to achieve real-time simulation and to avoid the need for embedding expensive CFD solver in the customized engine specific XactLIFE systems. As an integral part of the engineering software, XactLIFE needs to provide the users with quantitative modeling uncertainties for the developed surrogate models. Toward this goal, an efficient methodology for uncertainty quantification and propagation for surrogate modeling framework will be developed by the research team from the University of Waterloo in close collaboration with the technical experts from the industrial partner. Uncertainty quantification will provide confidence levels on the reliability of the prediction results for aging engine health and provide more accurate predictive maintenance recommendations leading to lower emissions, reduction in wastage of spare parts by taking optimal corrective actions. This research project will significantly impact the capability and range of application for the XactLIFE system in terms of its efficient and cost-effective applications to various industries for predictive maintenance of engine components. The success of this project will enable the industrial partner to create new source of revenue generation and reach out to new clientele.**

该项目的目标是为生成的代理模型开发用于不确定性量化和传播的概率框架，以模拟昂贵的计算流体动力学(CFD)解算器，该计算流体动力学(CFD)解算器并入由工业合作伙伴(LPTi-Life Forecast Technologies Inc.)内部开发的基于物理的生命周期管理专家系统(LCM-ES)框架XactLIFE(TM)。LPTI利用XactLIFE系统为燃气轮机发动机部件提供基于预测的生命周期管理服务。作为向客户提供的新产品的一部分，LPTi的目标是为具有预选关键部件的特定发动机定制XactLIFE系统。这将需要用代理模型替换XactLIFE系统中存在的CFD解算器，以实现实时模拟，并避免在定制的特定于发动机的XactLIFE系统中嵌入昂贵的CFD解算器。作为工程软件的组成部分，XactLIFE需要为用户提供所开发的代理模型的定量建模不确定度。为了实现这一目标，滑铁卢大学的研究团队将与行业合作伙伴的技术专家密切合作，开发一种用于替代建模框架的不确定性量化和传播的有效方法。不确定性量化将为老化发动机健康预测结果的可靠性提供置信度，并提供更准确的预测性维护建议，通过采取最佳纠正措施来降低排放，减少备件损耗。这项研究项目将对XactLIFE系统的能力和应用范围产生重大影响，因为XactLIFE系统在发动机部件预测性维护方面的高效和成本效益应用于各个行业。该项目的成功将使行业合作伙伴能够创造新的收入来源，并接触到新的客户。**