(554b) Optimization of Cleaning in Process Equipment Via CFD Studies

In the fine chemical, pharmaceutical, and food industries, a variety of methods are used to clean work surfaces between production runs. To reduce downtime and the quantities of cleaning agents that are used, there is a need to ensure that cleaning operations are completed quickly and efficiently. Common processes for cleaning vessels and industrial equipment include soaking, reflux, and washing with a spray or jet of a cleaning agent. To clean the inner surfaces of piping systems, fluids may be pumped through them. The choice of cleaning method largely depends on the nature of the residues and the mechanisms by which they can be removed. For example, residues may be removed by either mechanical force (shear stress) or dissolution into a flowing cleaning agent, often a solvent or detergent solution.

We report on the use of computational fluid dynamics (CFD) to model four examples of cleaning processes. Through simulations, we estimate the distributions of the forces exerted on surfaces and the mass transfer (dissolution) rates during soaking, reflux, impingement of a solvent jet, and pumping. All studies are three-dimensional and involve turbulent flows. For the single-phase flow systems (pumping and soaking with stirring), we use the Reynolds-averaged Navier-Stokes (RANS) equations coupled with a high Reynolds number k-ε turbulence model and standard wall function. To model an impinging jet and falling condensate film (cleaning by reflux), we use the volume of fluid (VOF) method to describe the gas-liquid flow together with large eddy simulation (LES) for turbulence modelling. In all studies, we consider the removal of a layer of acetaminophen (paracetamol) by water or methanol, examples of liquids in which acetaminophen has a low and high solubility, respectively. The dissolution and transport of acetaminophen is modelled through a passive scalar transport equation. We use acetaminophen as one example of a molecule with a representative solubility and diffusivity for the types of residues that are encountered in pharmaceutical manufacturing.

Sample computational results are shown in the accompanying figure for the studies of solvent flow in a pipe (a) and an impinging solvent jet (b). The piping system has two straight sections, with one joined to the middle of the other to model a case with a dead leg. Residue removal rates by dissolution (solute fluxes) are shown along a line on the side of the longer pipe section. Part (b) shows the solute fluxes at varying radial distances from the impingement point of a jet of water or methanol with a solid surface. In these systems, operating parameters such as temperature have several effects, for example on the solvent viscosity and residue solubility and diffusivity. In the case of a jet, the effects on the cleaning rate are complex, and CFD offers a way to estimate cleaning rates as a function of system parameters. Furthermore, CFD can be used to assess the locations of hard-to-clean areas within vessels and piping systems and evaluate options to improve cleaning. Use of CFD therefore appears to be promising for predictive modelling of cleaning processes to assist optimization and ensure successful completion (and validation).

The research leading to these results has received funding from Enterprise Ireland Innovation Partnership Programme (IP-2017-0573).

Fig. 1. Spatial distributions of contaminant flux from (a) the inner surface of a 2” diameter pipe into flowing methanol at 30 °C and 60 °C and (b) a flat wall with an impinging jet of methanol at 30 °C, 60 °C, and water at 30 °C. Inset in (a) shows the solvent velocity magnitude (2 m/s average inlet speed) in a cross-section of the pipe junction, solvent flows from the left to the top (right end is closed), and solute fluxes are evaluated along the dashed orange line. Inset in (b) shows a visualization of the shape of the solvent jet soon after impact with the surface.