Multi-level analysis and optimization for resource-efficient high temperature gas phase processes

Abstract

The targets of the Paris agreement for the mitigation of climate change require a drastic reduction of the per capita resource consumption in particular in the developed countries. New lifestyles, products and production processes are key measures in this context but the reduction of resource consumption cannot be achieved without a significant increase in resource productivity across all industrial sectors including the chemical industry. This increase in resource efficiency is achievable only through the analysis and optimization of each lengthscale: the screening of promising process units at a high hierarchical level and targeted measures for the improvement of the mass and energy efficiency at smaller lengthscales and lower hierarchical levels. At the same time, overall gains in efficiency through non-intuitive combinations of different process units are enabled at the plant level. It is evident, that systemic, multi-level approaches are required that take all these aspects into account in order to exploit the full process intensification potential. This dissertation addresses this multi-level problem at the reactor and plant levels for the example of high temperature catalytic gas phase processes. These processes include endothermic reactions like the reforming of methane and synthesis of hydrogen cyanide as prominent examples that are key intermediate processes of the chemical supply chain. Production of HCN serves as a case study in this thesis motivated by a collaboration with Evonik Industries but the reactor analysis and design methods of this work have been equally applied to steam reforming of methane. At first, a general literature overview combined with the methodological foundations in chapter 2 is provided. Following the introduction of the application background in chapter 3 this thesis is structured into three parts: the first two chapters address the reactor level and the third chapter focuses on the plant level of the process hierarchy. At the level of the single synthesis compartment in a bundle of multiple compartments in chapter 4, the underlying physical transport mechanisms for momentum and heat are analyzed in order to identify the reactor intensification potential. At this level the modeling of radiative heat transfer is emphasized. It is found that radiation accounts for approximately 20% of total heat transferred to the reacting fluid for dimensions similar to the industrial reactor design. Furthermore, radiation becomes the dominant mode of heat transfer for channel widths greater than 1 × 10−2 m. The heat transfer at this level can be enhanced through an increase in the surface emissivity of the catalytic wall material for example through integrating SiC in the ceramic material: with an increase in the surface emissivity from 0.45 to 0.90, an increase in the reactor outlet temperature by 30K can be achieved in the HCN synthesis example. Vice versa, decreasing the width or diameter of the single synthesis compartment by 1 − 2mm results in an increase in product yield by 2 − 3%. It is demonstrated that space time yield, yield and fuel efficiency are competing objectives at this scale which can hardly be prioritized without the overall process context at the plant level. At the next higher level of the tube bundle in chapter 5 the furnace design is optimized emphasizing the arrangement of the individual tubes of the tube bundle. To achieve this target a reduced two-dimensional model of the furnace is selected over a high fidelity CFD modeling of the three-dimensional furnace. It is shown that convective heat transfer is negligible unless a tube is placed directly within the main flow regime. Two key design parameters are identified for the optimal tube bundle design: inter-tube view factors that represent the shadowing between tubes and the hot flue gas emissivity. View factors are a good indicator of the performance of a bundle arrangement but do not suffice due to the impact of the flue gas emissivities: increasing inter-tube distances is beneficial as long as the gas layer emissivities are sufficiently high. If the tube shadowing is decreased beyond a certain tipping point, the overall bundle performance declines. For the HCN case study, the optimal inter-tube distance is identified as 0.052m which is a slight improvement in average product yield and in inter-tube standard deviation compared to the industrial benchmark scenario. Staggered bundle arrangements are more favorable for radiative heat transfer than aligned tube arrangements because they have lower total inter-tube view factors. The emphasis at the plant level is to identify the overall most resource-efficient – in terms of mass and energy consumption – production process for the case study of HCN. Stateof- the-art procedures for design of efficient chemical processes, however, are either limited in terms of problem size or have to solve the mass and energy integration consecutively which may lead to suboptimal process designs. For this reason the FluxMax approach which is a method for simultaneous process synthesis and heat integration was developed in collaboration with Dominik Schack and it is outlined in chapter 6. Through a discretization of the thermodynamic state space and a description of process units as stoichiometric reaction equations, a linear feasible region is created. The most resource efficient process for HCN production is then identified using multi-objective optimization through weighing of linear objective functions for atom efficiency, waste minimization, heating and total duty as well as variable cost minimization. Non-intuitive process alternatives such as the negligibility of column design for the HCN case study are identified with the FluxMax approach and it is shown that atom efficiency can be improved by 39.5% through recycling of the byproduct H2 to synthesize the reactant NH3. Furthermore, variable cost are reduced by 67.6 %. Nonetheless, it is demonstrated that no unique optimum exists – both optimal process designs result in higher overall duty requirements. Instead, careful weighing of the objectives for specific site conditions is required. The results of the individual levels show that a multi-level approach is essential in order to increase the resource efficiency of the chemical industry. The methods that are demonstrated in this dissertation have been successfully transferred to reactor design in methane steam reforming. The FluxMax approach is particularly versatile because it is level-independent and not limited to high temperature processes. As such it has been applied to process unit and plant design for methanol production. This thesis shows that high temperature processes of the chemical industry can contribute to an increase in resource productivity but the global targets can only be achieved as a combined effort of all industries, individuals and societies as a whole.