Dylan Rubini

This paper introduces a new platform to accelerate the modeling of complex aerothermochemical interactions in new turbomachines, called turbo-reactors, to decarbonize chemical processes. While previous work has aerothermally demonstrated the potential to decarbonize the heat input to the reaction, optimizing the reaction efficiency has been a challenge. This is because measuring reaction performance with aerochemical simulations is computationally prohibitive due to the uniquely complex aerodynamics and chemistry within turbomachines. To address this, we introduce a new multifidelity machine-learning-accelerated methodology, called ChemZIP, to mitigate this bottleneck. Although data-driven acceleration methodologies exist for combustion, modeling reactive flows along the bladed path of a turbomachine poses new challenges. This has led to a novel training data generation process. Our approach allows rich dynamic responses of the chemical system to be embedded into the training dataset at a fraction of the cost of reacting flow simulations. The resulting high-dimensional composition vector is then compressed onto a low-dimensional basis using an autoencoder-like neural network. This data-driven method is inspired by flamelet-generated manifolds but is more universal. Verification against 1,000 unseen one-dimensional test conditions results in an R2 score exceeding 95% across all quantities of interest. Following this, ChemZIP is coupled into a fully-fledged turbulent computational fluid dynamics solver. For a set of process-relevant three-dimensional configurations entirely different from the training data, the predictive accuracy of the thermochemical state remains within 10% of an industry-standard solver (Fluent). Crucially, this is possible while achieving a convergence to steady state almost 50 times faster, even for a small mechanism. Therefore, numerical computations are sufficiently fast that aerothermochemical optimization is now feasible for the first time in the design cycle of the turbo-reactor.

Hard-to-abate industrial processes, such as petrochemicals and cement production, have long been considered technically challenging to decarbonize. In response to the urgent demand to eliminate industrial CO2 emissions, a new class of energy-imparting turbomachines has been developed. These devices aim to convert mechanical into internal energy instead of pressurizing the gas, which enables high-temperature gas heating (up to 1700 ∘C) for a variety of applications. This article is organized into three parts. First, this article aims to demonstrate the capabilities of the novel, high-capacity, customizable, repeating-stage axial turbo-heater architecture for a hydrocarbon cracking example application. The study presents the new design requirements and working principles of this energy-imparting concept. The radically different objectives compared to a compressor enable ultra-high loading stage designs by avoiding the stability and efficiency constraints imposed on compressors. Within this new design space, the turbo-heater is able to achieve a loading coefficient ψ ≥ 4.0. Second, detailed numerical simulations of a multistage axial turbo-reactor with various vaneless space lengths are conducted. This work conclusively demonstrates the robustness and flexibility of the aerodynamic design despite employing a uniform blade design. It is emphasized that the concept is tolerant to unsteady interstage turbulent disturbances, enabling nominal work input conditions to be achieved even for the most compact arrangements. Finally, having confirmed that the aerothermal restrictions on the vaneless space length can be removed, the designer is free to tailor the design to optimize the chemical reaction by (1) tailoring the residence time distribution to improve the yield and coking rate, (2) homogenizing reaction progress by mixing-out concentration gradients, and (3) adjusting the rotational speed to account for variations in the reaction dynamics for different feedstocks.

This paper presents a revolutionary turbomachinery concept, referred to as the turbo-reactor, which has the potential to replace gas-fired radiant furnaces and decarbonise a wide range of hard-to-abate, high-temperature endothermic chemical reaction processes. Although previous studies by the authors have confirmed the feasibility of using a turbo-reactor for steam cracking reactions, the numerical investigation presented in this work broadens the scope of potential applications for the machine to a variety of energy-intensive chemical processes, including those used for hydrogen production. This step change in technology could be the catalyst needed to enable rapid scale-up of low-carbon hydrogen technology. The innovative design of the turbo-reactor is fundamentally based on converting all of the imparted mechanical energy into internal energy, rather than pressure. This enables temperatures of up to 1,700°C to be achieved within an axial length on the order of one metre, resulting in an increase in the power density of 50 to 1,000 times compared to a surface heat exchanger. This paper presents the first comprehensive analysis of the turbo-reactor’s robustness and controllability across a broad spectrum of feeds, chemical reaction stages, Mach number regimes, and operating points, conclusively demonstrating the feasibility of a universal stage design strategy for repeatedly imparting and dissipating energy for various endothermic reaction processes.

Decarbonizing highly energy-intensive industrial processes is imperative if nations are to comply with anthropogenic greenhouse gas emissions targets by 2050. This is a significant challenge for high-temperature industrial processes, such as hydrocarbon cracking, and there have been limited developments thus far. The novel concept presented in this study aims to replace the radiant section of a hydrocarbon cracking plant with a novel turboreactor. This is one of the first major and potentially successful attempts at decarbonizing the petrochemical industry. Rather than using heat from the combustion of natural gas, the novel turboreactor can be driven by an electric motor powered by renewable electricity. Switching the fundamental energy transfer mechanism from surface heat exchange to mechanical energy transfer significantly increases the exergy efficiency of the process. Theoretical analysis and numerical simulations show that the ultrahigh aerodynamic loading rotor is able to impart substantial mechanical energy into the feedstock without excess temperature difference and metal temperature magnitude. A complex shockwave system then transforms the kinetic energy into internal energy over an extremely short distance. The version of the turboreactor developed and presented in this study uses a single rotor row, in which a multistage environment is achieved regeneratively by guiding the flow through a toroidal-shaped vaneless space. This configuration leads to a reduction in reactor volume by more than two orders of magnitude compared with a conventional furnace. A significantly lower wall surface temperature, supersonic gas velocities, and a shorter primary gas path enable a controlled reduction in the residence time for chemical reactions, which optimizes the yield. For the same reasons, the conditions for coke deposition on the turboreactor surfaces are unfavorable, leading to an increase in plant availability. This study demonstrates that the mechanical work input into the feedstock can be dissipated through an intense turbulent mixing process, which maintains an ideal and controlled pressure level for cracking.

A new class of turbomachines, called turbo-reactors, have emerged to decarbonize high-temperature chemical processes. These applications unlock a new aerothermochemical design space for turbomachines. This paper explains how the turbo-reactor has the potential to be a “chemically tuned” device. In contradiction with conventional design wisdom, this can be achieved by balancing entropy generation within the flow against heat absorption by the reaction. This enables the “design” of a reaction-efficient temperature profile. To do this, it is necessary to understand and quantify the distribution of the entropic and isentropic mechanisms responsible for energy conversion. This paper uses high- and low-fidelity computations to decompose the energy conversion process into mechanisms based on a spatial decomposition. A range of Mach and Reynolds number regimes are studied, as well as a multistage configuration without vaneless space. The energy conversion breakdown analysis indicates that the entropic energy conversion dominates over the isentropic component with a contribution of 65%. The dominant source of entropy production is viscous dissipation generated by the thick diffuser trailing edge, accounting for 25% of the total. The shock system provides 20% of the energy conversion, almost entirely due to reversible pressure rise rather than entropy production. The energy conversion coefficient is independent of Reynolds number over engine-relevant conditions, whereas Mach effects are more significant. Across the Mach numbers range 1.1 to 1.5, the energy conversion coefficient increases by 20%. This is lower than expected as a result of the opposing effects of reversible and irreversible energy conversion contributions.