(317b) Dynamic Optimization of a Reaction-Absorption Process for Low-Pressure Ammonia Synthesis

Ammonia is the basis for the production of fertilizers, which are necessary for improving crop yields and feeding the growing global population [1]. More recently, ammonia has also gained attention as a hydrogen and energy carrier. Ammonia production mostly uses the well-established and energy-intensive Haber-Bosch (HB) process, which has been in use for over a century. Green ammonia is produced using renewable energy sources such as wind or solar power, instead of fossil fuels like natural gas [2]. However, green ammonia production is still in the early stages of development, with most projects focused on feasibility studies for plant construction. Additionally, producing ammonia from renewable resources is currently more expensive than producing traditional ammonia from fossil fuels, posing a significant technical challenge in reducing costs and competing with cheaper alternatives [3].

Selective absorption of ammonia immediately after synthesis can lower the cost of the process by using metal halides like calcium chloride to replace traditional ammonia condensation [4]. The low-pressure process requires less expensive equipment. Moreover, operating at lower pressure reduces the risk of equipment failure and ammonia leaks [5]. However, the single-pass conversion is low at a reduced operating pressure. Therefore, optimizing the design and operation of the low-pressure alternative is necessary to increase the ammonia production rate and make it economically attractive. Optimization studies have been performed on the ammonia synthesis reactor, but most only consider the traditional HB process in a steady-state environment [6]. Optimal design of the absorption-based synthesis process has also been considered [7]. However, optimization of the synthesis reactor in a dynamic environment and of the overall synthesis loop, including both reaction and separation, has received little attention. Dynamic operation of such reactors is essential to provide flexibility in the presence of intermittent renewable energy sources.

The goal of this study is to explore the optimal dynamic operation of the reaction-absorption process for ammonia synthesis at low pressure. The process configuration includes a three-bed autothermal reactor with direct cooling through quenching, followed by three heat exchangers and an absorption column. Six process variables are simultaneously manipulated, including feed flow rate, feed temperature, three quenching flow rates, and recycle flow rate. The ammonia synthesis reactor's individual beds are modeled as adiabatic, pseudo-homogeneous plug flow reactors, with temperature and concentration gradients occurring in the axial direction. The absorber column is described using an axial dispersed plug flow model with gas velocity varying along the axial direction. The model of the entire ammonia synthesis loop consists of a set of partial differential equations, an ordinary differential equation, and algebraic equations. A dynamic optimization problem was formulated where the objective is the maximization of the ammonia production rate subject to the system model constraints.

The optimal profiles obtained in the present study show a consistent pattern over time that can be divided into three distinct parts. The first part corresponds to a transient phase, characterized by significant changes in the profiles. The second part represents a non-transient phase where an optimal steady state is maintained. Finally, the third part marks another transient phase where the profiles deviate from the steady-state value. The state variables, ammonia mass fraction and absorption rate follow a similar pattern. This behavior is referred to as the turnpike property [8]. This property, first identified in econometrics, suggests that the solution of an optimal control problem should spend most of its time near a steady state, and in infinite horizon, the solution should converge to that steady state. In the context of the ammonia synthesis loop, the turnpike property implies that keeping the manipulated variables close to an optimal steady-state value for most of the time horizon can result in a more cost-efficient trajectory, while maintaining a high ammonia production rate. Future work aims to include the desorption of ammonia and simultaneous heating and absorption, as well as the analysis of the turnpike property for control and optimization purposes.

[1] Barona, A., Etxebarria, B., Aleksanyan, A., Gallastegui, G., Rojo, N., and Diaz-Tena, E. (2018). A unique historical case to understand the present sustainable development. Science and Engineering Ethics, 24:261–274.

[2] Lee, B., Winter, L. R., Lee, H., Lim, D., Lim, H., and Elimelech, M. (2022). Pathways to a green ammonia future. ACS Energy Letters, 7(9):3032–3038.

[3] Salmon, N. and Bañares-Alcántara, R. (2021). Green ammonia as a spatial energy vector: a review. Sustainable Energy Fuels, 5:2814–2839.

[4] Malmali, M., Reese, M., McCormick, A. V., and Cussler, E. L. (2018). Converting wind energy to ammonia at lower pressure. ACS Sustainable Chemistry & Engineering, 6(1):827–834.

[5] Kale, M. J., Ojha, D. K., Biswas, S., Militti, J. I., McCormick, A. V., Schott, J. H., Dauenhauer, P. J., and Cussler, E. L. (2020). Optimizing ammonia separation via reactive absorption for sustainable ammonia synthesis. ACS Applied Energy Materials, 3(3):2576–2584.

[6] Gaines, L. D. (1977). Optimal temperatures for ammonia synthesis converters. Industrial & Engineering Chemistry Process Design and Development, 16(3):381–389.

[7] Palys, M.J., McCormick, A., Cussler, E.L., Daoutidis, P. (2018). Modeling and Optimal Design of Absorbent Enhanced Ammonia Synthesis. Processes, 6, 91.

[8] Trélat, E. and Zuazua, E. (2015). The turnpike property in finite-dimensional nonlinear optimal control. Journal of Differential Equations, 258(1):81–114.