GOALI/Collaborative Research: Control-Oriented Modeling and Predictive Control of High Efficiency Low-emission Natural Gas Engines

GOALI/协作研究：高效低排放天然气发动机的面向控制的建模和预测控制

• 批准号: 2302217
• 负责人: Javad Mohammadpour Velni
• 金额: $ 23万
• 依托单位: Clemson University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-10-01 至 2024-10-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2302217&HistoricalAwards=false
• 关键词: GOALI; Collaborative; Research; Control; Oriented

About 200 million internal combustion engines (ICEs) are produced in the world every year and used in energy, transport and service sectors. Furthermore, ICEs account for over 22% of the U.S. total energy consumption and produce the largest portion of CO2 greenhouse gas emissions in urban areas. Dual fuel natural gas (NG) engines in advanced low temperature combustion regimes represent the state-of-the-art ICE technology with some of the highest reported fuel conversion efficiencies and 25% lower CO2 emissions compared to conventional engines. However, achieving a robust and high-efficiency performance of these engines on a broad operational range using existing control technologies is not possible due to their highly nonlinear and uncertain dynamic behavior. This research aims at developing fundamental tools for dynamic modeling and control of nonlinear systems and applying them to high-efficiency low-emission advanced ICEs. The project will provide wide-ranging societal benefits through three major impact areas: first, by advancing research in nonlinear control systems, and mixing and reactive flow including combustion systems; second, by providing direct benefits for control of combustion engines, commonly used in power generation, automotive, locomotive, marine, oil and gas drilling, construction, utilities and manufacturing industries; and third, through educational and outreach activities delivered at industry sites, local communities and science fairs. This project is a collaborative effort between Michigan Technological University, University of Georgia, and the industry partner, Cummins Inc. The project intends to develop a suite of innovative control-oriented modeling and stochastic predictive control design tools to address control challenges for advanced dual fuel natural gas engines, as well as a broad range of other nonlinear and stochastic dynamic systems. The outcomes of this project result in six main components that include: (i) characterizing the dynamics of dual fuel NG engines in advanced combustion regimes, (ii) building the first physics-based control-oriented model for advanced dual fuel NG engines, (iii) developing new analytical tools for deriving models through the powerful fusion of machine learning and classical multivariate methods, (iv) providing solutions to fill the gaps between first-principles models and data-driven methods for estimating an accurate model, (v) bridging the gaps between parameter-varying systems and stochastic controls, and (vi) constructing, testing, and validating the combustion controllers for dual fuel NG engines. The outcomes from these six theoretical, modeling and experimental contributions will be generic dynamic modeling and predictive control design tools for nonlinear and stochastic industrial systems that are demonstrated on engine test-beds.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

全球每年生产约 2 亿台内燃机 (ICE)，用于能源、运输和服务行业。此外，内燃机消耗的能源占美国总能源消耗的 22% 以上，并产生了城市地区二氧化碳温室气体排放的最大部分。先进的低温燃烧状态下的双燃料天然气 (NG) 发动机代表了最先进的 ICE 技术，与传统发动机相比，其燃料转换效率最高，二氧化碳排放量低 25%。然而，由于这些发动机的高度非线性和不确定的动态行为，使用现有的控制技术在广泛的操作范围内实现稳健和高效的性能是不可能的。本研究旨在开发非线性系统动态建模和控制的基本工具，并将其应用于高效低排放的先进内燃机。该项目将通过三个主要影响领域提供广泛的社会效益：首先，推进非线性控制系统以及包括燃烧系统在内的混合和反应流的研究；其次，为通常用于发电、汽车、机车、船舶、石油和天然气钻探、建筑、公用事业和制造行业的内燃机控制提供直接的好处；第三，通过在工业现场、当地社区和科学博览会上开展教育和外展活动。该项目是密歇根理工大学、佐治亚大学和行业合作伙伴康明斯公司之间的合作成果。该项目旨在开发一套创新的面向控制的建模和随机预测控制设计工具，以解决先进双燃料的控制挑战天然气发动机，以及各种其他非线性和随机动态系统。该项目的成果包括六个主要部分，包括：（i）表征先进燃烧状态下双燃料天然气发动机的动力学特征，（ii）为先进双燃料天然气发动机构建第一个基于物理的控制导向模型，（ iii) 通过机器学习和经典多元方法的强大融合来开发新的分析工具来推导模型，(iv) 提供解决方案来填补第一原理模型和数据驱动方法之间的差距，以估计准确的模型，(v) 桥接这参数变化系统和随机控制之间的差距，以及 (vi) 构建、测试和验证双燃料天然气发动机的燃烧控制器。这六项理论、建模和实验贡献的成果将是用于非线性和随机工业系统的通用动态建模和预测控制设计工具，并在发动机试验台上进行演示。该奖项反映了 NSF 的法定使命，并被认为值得通过以下方式获得支持：使用基金会的智力价值和更广泛的影响审查标准进行评估。

项目成果

Control-Oriented Data-Driven and Physics-Based Modeling of Maximum Pressure Rise Rate in Reactivity Controlled Compression Ignition Engines

反应控制压燃发动机中最大压力上升率的面向控制的数据驱动和基于物理的建模

• DOI: 10.4271/03-16-06-0040
• 发表时间: 2023-01
• 期刊: SAE International Journal of Engines
• 影响因子: 1.2
• 作者: Irdmousa, Behrouz Khoshbakht;Basina, L. N.;Naber, Jeffrey;Mohammadpour Velni, Javad;Borhan, Hoseinali;Shahbakhti, Mahdi
• 通讯作者: Shahbakhti, Mahdi

Safe control of nonlinear systems in LPV framework using model-based reinforcement learning

使用基于模型的强化学习对 LPV 框架中的非线性系统进行安全控制

• DOI: 10.1080/00207179.2022.2029945
• 发表时间: 2022-01
• 期刊: International Journal of Control
• 影响因子: 2.1
• 作者: Bao, Yajie;Mohammadpour Velni, Javad
• 通讯作者: Mohammadpour Velni, Javad

Data-Driven Model Learning and Control of RCCI Engines based on Heat Release Rate

基于热释放率的 RCCI 发动机数据驱动模型学习和控制

• DOI: 10.1016/j.ifacol.2022.11.249
• 发表时间: 2022-01
• 期刊: IFAC-PapersOnLine
• 作者: Sitaraman, Radhika;Batool, Sadaf;Borhan, Hoseinali;Velni, Javad Mohammadpour;Naber, Jeffrey D.;Shahbakhti, Mahdi
• 通讯作者: Shahbakhti, Mahdi

Physics-guided and Energy-based Learning of Interconnected Systems: from Lagrangian to Port-Hamiltonian Systems

互连系统的物理引导和基于能量的学习：从拉格朗日系统到哈密尔顿港系统

• DOI: 10.1109/cdc51059.2022.9992803
• 发表时间: 2022-12
• 期刊: IEEE 61st Conference on Decision and Control (CDC
• 作者: Bao, Yajie;Thesma, Vaishnavi;Kelkar, Atul;Velni, Javad Mohammadpour
• 通讯作者: Velni, Javad Mohammadpour

Machine Learning-based Classification of Combustion Events in an RCCI Engine Using Heat Release Rate Shapes

使用热释放率形状对 RCCI 发动机中的燃烧事件进行基于机器学习的分类

• DOI: 10.1016/j.ifacol.2022.11.248
• 发表时间: 2022-01
• 期刊: IFAC-PapersOnLine
• 作者: Sitaraman, Radhika;Batool, Sadaf;Borhan, Hoseinali;Velni, Javad Mohammadpour;Naber, Jeffrey D.;Shahbakhti, Mahdi
• 通讯作者: Shahbakhti, Mahdi