BRITE Pivot: Enabling Efficient Fuel Cell Control Using Data-Driven Modeling

点击翻译按钮获取中文摘要

This award is funded in whole or in part under the American Rescue Plan Act of 2021 (Public Law 117-2).This Boosting Research Ideas for Transformative and Equitable Advances in Engineering (BRITE) Pivot award will fund research that makes the use of fuel cells as vehicle power sources a viable strategy for reducing carbon emissions from the transportation sector, thereby promoting the progress of science, advancing the national prosperity and welfare, and increasing the reliance on secure and environmentally friendly sources of energy. Fuel cells provide a method of electrifying vehicles that is well suited to medium and heavy-duty vehicles, since they can be refueled and provide a long driving range but can operate with a much higher efficiency and reduced or negligible emissions compared to petroleum-based engines. The highly variable power demands typical of automotive applications make proper control of the dynamics of a fuel cell stack challenging. New tools are needed to ensure that internal variations in fuel or air flow, temperature, or humidity can be understood, modeled, and regulated to ensure sustained performance and optimized longevity. This project will address this need by infusing the study of fuel cell dynamics with data-driven modeling techniques that enable real-time control of the internal state of a fuel cell stack. The principal investigator will leverage past research experience in conventional engine modeling and control to achieve a critical leap forward of a complex technology of significant sociotechnical importance. New online courses on the hydrogen energy economy will help broaden participation in STEM of students from currently underrepresented groups and the institutions that serve them. A pilot mentoring program will connect students with disciplinary experts in automotive and transportation systems. This research aims to make fundamental contributions to the modeling and control of internal fuel cell stack dynamics with emphasis on the impact, detection, and mitigation of fuel impurities. It will achieve this outcome by developing data-driven, physics-based, reduced-order models that bridge the gap between calibration-intensive, lumped-parameter models and computationally demanding gas dynamics models, neither of which allow for real-time control of the fuel cell state. The project will combine foundational experiments with theoretical advances in the integration of multi-dimensional models with artificial neural networks that significantly lower the computational cost while capturing important nonlinearities, reaction rates, and aging effects. Nonlinear model-predictive control techniques that leverage this model structure will be explored and paired with a model-based fuel impurity detection methodology. The capabilities of the final control approach will be demonstrated experimentally at steady-state and transient conditions illustrative of those encountered in automotive drive cycles.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.