(708d) A Computational Study of Electrically Heated Reactors

Most chemical transformations and production require heating, currently based on burning fossil fuels, which cause significant CO₂ emissions. About half of the CO₂ emissions in the chemical industry are related to fuel combustion for heat supply. Recently, a major effort is underway in using renewable electricity in the chemical industry. Out of various electrification routes, Joule heating is at the forefront of scientific exploration. Electric heating allows the direct heating of the catalyst, keeping the catalytic sites at a higher temperature than the gas phase resulting in reduced side reactions, better catalyst utilization, and higher productivity. Thus, electric heating could allow modular and compact reactors. Electric heating also provides start-up and shut-down time in seconds, making it a suitable candidate for intermittent usage. Several reactor configurations are being explored for their potential as electric reactors. However, realizing the full potential of electric reactors remains a challenge. Most of the investigations are still at the lab scale and require significant efforts to develop and deploy electric reactors at the industrial scale. At present, most investigations are experimental and use a trial-and-error-based approach, which is inefficient and cannot explore a vast design space. In this regard, high-fidelity and high-throughput computer models that can test a vast design space and operating regime are imperative. Such tools can form the basis for developing rules of thumb for emerging electric reactors. This work performs first-principles simulations of electrically heated reactors by solving a coupled thermo-electric problem. These first-principles simulations provide insights into the temperature profile within the reactor obtained during the electric heating. Moreover, a surrogate model is developed and assessed against the 3D simulation predictions for a wide range of operating conditions. We show that the surrogate model allows accurate predictions of the reactor temperature at a fraction of the 3D simulation cost.