(451c) Supermode-Based Temperature Control of Inductively Heated Chemical Reactors

The chemicals industry accounts for about 2 billion metric tons of CO₂ emissions every year, which is about 5% of the total global greenhouse gas emissions. Most of those emissions come from the need to heat volumetric media for processing materials, syngas generation, and gas separations. Compared to fossil fuel heating, electrified heating has a lower carbon footprint and brings forth process intensification benefits such as better thermal transfer, faster start-ups, and higher temperature control precision. Though there are several electric heating techniques, induction heating has particular promise for chemical processing as it is maturely used in the metallurgy industry to volumetrically heat metal efficiently at megawatt power levels. Magnetic induction involves injecting alternating current (AC) into a helical coil to create an alternating magnetic field, which wirelessly couples to a susceptor to generate joule heating.

In a recent previous study, our group demonstrated that with proper co-design of the reactor geometry, material, and power electronics, induction heating can achieve power-to-heat efficiencies close to unity while volumetrically delivering high grade heat. We experimentally inductively heated a SiSiC ceramic susceptor to perform the endothermic reverse water-gas shift (RWGS) reaction with a marginal experimental efficiency of 86.6%, minimal radial temperature gradients, and close to equilibrium conversions of CO₂ for various gas flow rates. However, those homogeneously heated susceptor reactor systems featured axial temperature gradients that limited the total reactor size. These gradients were caused by thermal conduction losses at the susceptor ends, convection cooling due to the inlet gas flow, and heat consumption from the endothermic reaction, and they became more pronounced under dynamic operating conditions such as higher input powers, varying input gas flow rates and/or molar ratios, and lifetime degradation of catalysts. As such, a system that can make real time axial heating profile adjustments can minimize these gradients and enable high conversion across a reactor’s runtime.

To address this, we propose a “supermode” inductively heated chemical reactor. The concept of a supermode arises from the mutual interactions among coupled resonators. In classical mechanics, this refers to the collective oscillation of coupled harmonic oscillators. In optics, it describes the superposition of electromagnetic fields in coupled waveguides. We extend this idea to inductive heating by designing a reactor with multiple mutually coupled coils, which support supermodes of magnetic field distributions. By selectively choosing and exciting different supermodes, we can generate arbitrary and spatially controlled heating profiles, enabling us to tailor the axial temperature distribution and ultimately achieve higher reaction conversion. The supermode scheme has several advantages compared to other “zone control” induction heating schemes. Such schemes require multiple amplifiers with complex current sensing controls to avoid coil coupling, bottlenecked by the speed of sensing. The supermode concept requires only one amplifier to drive the entire system with simple controls that require sensing on the order of the thermal time constant of the system. Other zone control systems can only turn on and off certain “zones”, which can overshoot and undershoot temperatures leading to oscillations, coarse heating profiles, thermal shocks that can stress the catalysts and susceptor, and overall lower energy efficiency. In contrast, the proposed supermode system heats all zones simultaneously with a more continuous profile yet is still capable of on/off control. Lastly, a supermode reactor can take advantage of the mutual coupling of neighboring coils to produce higher order heating profiles.

In our design, each coil is loaded with a tunable series capacitance that can change its AC current magnitude. By analyzing a circuit model, we can determine the range of capacitance values set by the desired heating profiles with a set of closed-form expressions. We use multiphysics software (i.e. COMSOL) to validate the circuit model and the temperature profile given a magnetic field profile. We experimentally demonstrate the supermode heating scheme with a 7” long, 2” diameter additively manufactured open-cell SiSiC foam packed with K₂CO₃/Al₂O₃ catalysts to perform RWGS. The reaction system comprises 8 coils, loaded by discrete capacitances switched with relays, supplied by a linear power amplifier capable of delivering power at 13.56 MHz of up to 1 kW. We build a model predictive controller (MPC) to correct for external disturbances using a multipoint fiber optic thermocouple to measure temperature along the axial profile with a microcontroller that updates the magnitude of currents in each coil. We demonstrate that we can regulate the axial temperature for the following disturbances: (1) fluctuations of inlet flow rate (2) fluctuations of input gas molar ratio of H₂ and CO₂ (3) long term degradation and deactivation of catalysts and (4) user input changes to the steady state temperature setpoint. Additionally, the MPC has a response time of 200 milliseconds, which is much lower than the time constant of the system. We showcase the conversion enhancement normalized to a setpoint temperature by going from a conventional inductively heated reactor to a supermode-based reactor.