Real-time visualization of CO2 electrolysis

该项目将推动电化学反应器（电解槽）的发展，通过将二氧化碳捕获解决方案转化为社会广泛使用的产品，有可能减少温室气体排放。项目团队包括：UBC大学的Curtis Berlinguette教授，他是二氧化碳捕获和转化研究的全球领导者；UdeM的Yoshua Bengio教授，机器学习算法开发的先驱；Carbon Engineering是一家全球领先的二氧化碳捕集公司，总部位于不列颠哥伦比亚省斯夸米什。电化学CO2转换是一种很有前途的策略，可以利用大气或排放的CO2将可再生能源储存在碳基化学品（例如CO）中。电解槽大部分是不锈钢外壳的封闭系统。因此，可以测量输入和输出流，但不能直接观察到反应器内发生的化学和物理现象，并且对其了解甚少。Berlinguette集团正在开发一种创新的成像技术，该技术支持该项目，使科学家首次能够可视化电解槽内运行的过程。这个可视化工具可以产生巨大的数据集，这些数据集太大而无法有效地手工处理；通过这个项目，Bengio集团将利用机器学习（ML）对捕获的数据进行建模，以便更好地分析数据，并为底层流程的优化提供指导。在这个项目中，二氧化碳电解的实时可视化将有助于开发基本的流体流动模型，为项目团队以及更广泛的科学和工业社区提供有价值的见解。我们将利用项目经验来确定提高转化率和能源效率的反应堆材料、反应堆配置和运行条件。这些进步是未来商业规模反应堆部署所需要的。Carbon Engineering将为该项目带来工业视角，同时深入了解机器学习如何对他们当前的工作流程有用。该项目还将培训加拿大从事气候科学职业的年轻科学家和工程师，并为雄心勃勃的温室气体减排机会提供信息。

This project will advance electrochemical reactors (electrolysers) that have potential to mitigate greenhouse gas emissions by converting CO2 capture solutions into products used widely by society. The project team includes: Prof. Curtis Berlinguette at UBC, a world leader in CO2 capture and conversion research; Prof. Yoshua Bengio at UdeM, a pioneer in machine learning algorithm development; and Carbon Engineering, a global leading CO2 capture company headquartered in Squamish, BC. Electrochemical CO2 conversion is a promising strategy to enable renewable energy to be stored in carbon-based chemicals (e.g., CO) using atmospheric or emitted CO2. Electrolysers are largely closed systems with stainless steel housing. Consequently, input and output streams can be measured, but the chemical and physical phenomena occurring inside the reactor cannot be directly observed and are poorly understood. The Berlinguette Group is developing an innovative imaging technique that underpins this project and allows scientists to visualize the processes inside an operating electrolyser for the first time. This visualization tool can produce enormous data sets that are too large to be efficiently processed manually; through this project the Bengio Group will leverage machine learning (ML) to model the captured data in order to better analyze it and provide guidance for the optimization of the underlying process. Real-time visualization of CO2 electrolysis during this project will enable development of fundamental fluid flow models that provide valuable insights for the project team as well as the broader scientific and industrial communities. We will leverage project learnings to identify reactor materials, reactor configurations, and operating conditions that enhance conversion rates and energy efficiency. Such advancements are needed for future reactor deployment at commercial scale. Carbon Engineering will bring an industrial perspective to the project while gaining insight into how ML may be useful to their current workflows. The project will also train young scientists and engineers in Canada for careers in climate science and inform ambitious GHG mitigation opportunities.