Continuous Polymerization in Modular, Intelligent Reactors Resistant to the Formation of Deposits
The chemical industry is one of the most energy-intensive production sectors in Germany, and its production processes still offer considerable potential for energy savings. While the production of petrochemical raw materials and basic chemicals is already carried out in highly energy-efficient continuous processes, the production of pharmaceuticals, fine and specialty chemicals still generally uses batch processes with low energy efficiency in multi-product plants. As part of the ENPRO Initiative I and II, modular and flexible plant concepts have been and are being developed in order to be able to use the advantages of a continuous production mode for the fabrication of smaller and special chemical products. A major obstacle to the rapid implementation of these new concepts is the occurrence of fouling and deposits, which can severely disrupt continuous operation. In the joint project KoPPonA 2.0, the implementation of continuous process concepts for various polymer specialties which are particularly susceptible to the formation of deposits is to be promoted. Therefore, plant operators, apparatus manufacturers, sensor manufacturers, material scientists and process engineers work closely together to elucidate the causes of coating formation and to ensure the operation of continuous plants through innovative approaches in apparatus design, surface modification and reaction control.
Figure 1: The ENPRO initiative funded by the German Ministry of Economy and Climate Action and partners in the KoPPonA 2.0 project
The Institute of Multiphase Flows uses computational fluid dynamics (CFD) to investigate the processes leading to fouling. Due to the high Schmidt number of the problem, a high resolution is required to eliminate numerical diffusion. Different computational methods, i.e., finite element method (FEM), finite volume method (FVM), and Lattice Boltzmann method (LBM) are compared within the project to benchmark computational performance at high grid resolutions. Furthermore, a Euler-Lagrangian method is used in FVM to predict homogeneous and heterogeneous polymer fouling in structured reactors. Together with the Ruhr-University Bochum, the results are used to derive a compartment model of the fluid dynamics to significantly reduce computational effort.
Preventing Fouling by means of Pre-Mixing
Many continuous processes rely on a pre-mixing stage to achieve ideal mixing before the reagents enter the reactor stage. The pre-mixer is usually by orders of magnitude smaller than the continuous reactor, i.e., a milli- or micro-mixer. The mixing on molecular scale depends on the complex interplay between fluid dynamics, mass transfer and chemistry. Conventional milli- and micro mixers are investigated by means of
Direct Numerical Simulations (Figure 2 A),
the novel Imaging UV/Vis Spectroscopy to capture transient and reactive multi-component systems in 2D (Figure 2 B) and
Confocal Laser Scanning Microscopy using Laser-Induced Fluorescence (CLSM-LIF) to record stationary 3D concentration fields (Figure 2 C).
Figure 2: (A) DNS of a split-and-recombine mixer, (B) imaging UV/Vis spectroscopy on a split-and-recombine mixer, (C) CLSM-LIF on a split-and-recombine mixer.
In CFD simulations the grid resolution needs to account for the length scale of mass transfer in miscible liquid-liquid systems (high Schmidt number problem). One main focus point of this research lies in investigating the length scale of non-reactive mixing and reactive mixing, respectively.
Figure 3: Competitive reaction of protons in a T-micro mixer (left): Momentum and reaction equations are grid dependent at high Schmidt numbers (Sc = 1000, Re = 200).
The grid resolution ∆S/d (normalized by hydraulic diameter) heavily influences the solution of the governing equations. Figure 3 shows how the momentum solution converges at relatively low resolutions (for low Re), however the reaction solution requires much larger resolutions to converge.
Numerical Modeling of Deposit Formation
Deposit Formation (Fouling) in polymer solutions is driven by two mechanisms. Homogeneous fouling describes the growth of deposits on surfaces due to increasing polymer chain length and solution viscosity (auto-acceleration) in near-wall regions. Figure 3 shows the local increase of viscosity in regions of large residence times (i.e., vortex structures). Eventually, the chain length is sufficiently long so that gels or solids forms at the wall, leading to the plugging of the reactor. This behavior is modeled in CFD by means of a solution-viscosity approach dependent on the polymer reaction yields. This allows for optimization of geometry and operating conditions to delay and prevent cleaning intervals.
Figure 4: Homogeneous (left) and heterogeneous (right) fouling mechanisms on walls.
As the second mechanism, heterogeneous fouling (Figure 4 right) occurs through precipitation of solid polymers from the solution. These polymers can grow in size, coalesce and accumulate at surfaces. The precipitation is modeled by an Euler-Lagrangian approach. The reaction rate of the continuous (Eulerian) phase locally creates discrete (Lagrangian) particles. Both the phases are coupled to form a model predicting heterogeneous fouling locally within the reactor, as shown in Figure 5. Due to the high computational effort, only stationary simulations are feasible, and time cannot be resolved. However, the tendency of the simulation to numerically diverge indicates that fouling is likely to occur at the respective operating conditions.
Figure 5: Euler-Lagrangian particle trajectories in the reactor domain (top), particle accretion (middle) on heated reactor walls computed with the Euler-Lagrangian model and resulting pressure drop (bottom) of the continuous phase on the reactor mid-section.
Modeling for example the particle-wall interaction, the accumulation of heterogeneous deposits is investigated locally to optimize the reactor geometry and allow for long operating times of the system.