Different operational modes, various scales and complex phenomena make the design of a chemical process a challenging task. Besides conducting basic lab experiments and deriving fundamental kinetic and thermodynamic models, a crucial task within the entire process design is the synthesis of an optimal reactornetwork constituting the core of a chemical process. However, instead of directly up-scaling the process to large devices, it is wise to investigate process characteristics on miniplant-scale. For an existing miniplant for the hydroformylation of 1-dodecene using a rhodium catalyst and a thermomorphic solvent system for catalyst recovery, two optimized reactor designs are derived. Suitable reactor-networks were synthesized by applying the Flux Profile Analysis approach introduced in Kaiser et al. (2017). The combination of a first reactor with dynamic/distributed control options and a subsequent backmixed CSTR arose to be the most promising configurations. The technical design under miniplant conditions were carried out for two possible realizations of this network, namely (i) a continuous flow reactor and (ii) a periodically operated semibatch reactor, both followed by the existing CSTR which was originally operated in the miniplant. An optimization of the two optimal reactor configurations within an overall process including a Page 1 of 53 ACS Paragon Plus Environment Industrial & Engineering Chemistry Research liquid-liquid phase separation for catalyst recovery and a distillation column for separating the solvents and reactant evinced a selectivity w.r.t. the linear aldehyde around 94 % and a conversion around 98 %. This is a large improvement of the process performance of 24 % linear aldehyde selectivity and 40 % conversion when using the existing CSTR.

A lingering issue with the hydroformylation of long-chain alkenes is the cost of catalyst leaching. One effective method to recover homogeneous catalysts is the use of thermomorphic solvent systems (TMS). However, catalyst leaching is still too high using the current solvents DMF and decane, limiting economic feasibility. This work presents extraction as a possible method for intensifying catalyst recovery when using a TMS for the hydroformylation of 1-dodecene. A thermodynamic model for determining the LLE of the solvent system and for catalyst leaching is developed for implementation within a process-wide optimization problem. Using this model, the optimal reactor design with an integrated downstream separation including the catalyst loss can be investigated in more detail. It is shown that in this process the reactor design strongly depends on catalyst recovery and that by using the proposed extraction cascade the process becomes economically viable and more robust in regards to reactor performance.

Within the framework of elementary process functions (Freund and Sundmacher (2008)) an approach is developed to derive reactor-network candidates from the solution of a dynamic optimization of a batch process scheme by analyzing its optimal mass and energy control fluxes. Thereby, any characteristics of the reaction progress can be identified, e.g. benefits from mixing, back-mixing, recycling, heating, cooling, etc.The approach is used to (i) determine the attainable region for the modified, isothermal van-de-Vusse reaction, which matches literature results; and (ii) synthesize reactor-network candidates for the standard, non-isothermal van-de-Vusse reaction, which gives new insights compared to previous results from literature using superstructure optimization approaches. The results indicate how this approach can be used to determine the attainable region of a process and to rationally select candidates for detailed reactor design with, e.g. superstructure optimization. It further closes the gap between dynamic batch optimization and continuous reactor-network synthesis.

No abstract

Volumetric mass transfer coefficients and Bodenstein numbers (Bo) for the elongated bubble flow regime in horizontal helically-coiled tube reactors are reported using two different measurement techniques (oxygen optodes, and optical observation of an oxygen-sensitive dye). Additionally, the gas-liquid mass transfer and the residence time behavior of the two-phase flow were described with a 3D Computational Fluid Dynamics (CFD) model, and also with a onedimensional two-phase model. For this study, 16 cases involving different gas and liquid volumetric flow rates were employed to generate air-water flows through two helically coiled tubes with curvature ratios of δ 1 = 0.093 and δ 2 = 0.3, respectively. The superficial gas and liquid Reynolds numbers (Re s,G and Re s,L) and the gas holdup (ϵ G) are varied from 494 to 2483, from 1456 to 2713, and from 0.46 to 0.81, respectively. The mass transfer measurements show an increasing gas-liquid mass transfer rate with increasing superficial velocity of the liquid-phase and Re s,G. The Bodenstein number decreases with increasing gas-phase Reynolds number and increases with increasing superficial velocity of the liquid-phase. Correlations describing the mass transfer and backmixing behavior are proposed. The CFD results are in excellent agreement with the experimental data. With the 1D two-phase model it is possible to describe the residence time behavior of the two-phase flow through the helically coiled tube.