Heat Transfer in a Plain Jacket of a Pilot Scale Stirred Tank Reactor

Abstract

Recent literature on thermal runaway reaction research in vessels, such as that conducted by Rudniak et al. [2011], still assumes a constant jacket temperature. In this investigation, three-dimensional steady state CFD simulations are performed for the plain jacket of a pilot scale vessel, which predict that the jacket temperature can vary by tens of degrees Celsius across different parts of the jacket even in reactors under a third of a metre in diameter, and that the distribution of heat transfer coefficients is strongly dependent on the flow path. The CFD output values are compared with experimental data of temperature measurements and with the use of correlations to predict heat transfer coefficients from the experimental data. The commercial code Ansys CFX Version 15 has been used to simulate the flow of the commercial heat transfer oil "DW-Therm" through a plain jacket of a pilot scale stirred tank reactor that is boiling water inside the vessel, to approximate a constant process temperature.

CFD modelling of conjugate heat transfer in a pilot‐scale unbaffled stirred tank reactor with a plain jacket

“…The fully‐conjugate value of is the largest at 116.6 W m −2 K −1 and the semi‐conjugate values are approximately 13 % smaller. In addition, an OHTC value from an earlier CFD investigation for flow in the jacket and boiling of methanol in the vessel resulted in a slightly lower value. In this case the boiling heat transfer coefficient (calculated from the Cooper correlation) will be much higher than those predicted values from the CFD simulation on the process side of the vessel and accordingly will have a smaller thermal resistance.…”

“…For the lower wall in Figure a, the mean jacket temperature was set to 78.2 °C, based on the experimental values. The heat transfer coefficient was set to 220 W m −2 K −1 , which is based on the values obtained from a previous CFD investigation of the flow in the jacket . The upper wall in Figure a is exposed to an ambient temperature of 20 °C.…”

“…However, the medium mesh was used for the simulations and analysis of results presented in this paper. In the jacket, four different meshes consisting of 1.75 × 10 6 , 3.25 × 10 6 , 6.16 × 10 6 , and 9.22 × 10 6 hexahedral elements were made . The minimum width of the computational cells near the wall was chosen to equate the value less than 2, to resolve the flow near the wall.…”

“…Yapici and Basturk carried out semi‐conjugate (coupled process and wall) heat transfer simulations in an unbaffled hemispherical agitated vessel, not representative of industrial STRs, using the temperature and heat transfer coefficient as boundary conditions at the outer surface of the vessel wall. Our previous CFD study of flow and heat transfer in the plain jacket of a pilot‐scale 25 L STR found that the temperature distribution of the flow could vary by tens of degrees, and non‐uniform distributions of heat transfer coefficient and shear stress on the inner wall of the jacket, which could result in uneven heating of the process fluid . This necessitates coupling of the jacket and the vessel in the CFD modelling using a conjugate heat transfer approach, which can be very challenging.…”

“…Our previous CFD study of flow and heat transfer in the plain jacket of a pilot-scale 25 L STR found that the temperature distribution of the flow could vary by tens of degrees, and nonuniform distributions of heat transfer coefficient and shear stress on the inner wall of the jacket, which could result in uneven heating of the process fluid. [12] This necessitates coupling of the jacket and the vessel in the CFD modelling using a conjugate heat transfer approach, which can be very challenging. However, to the best of authors' knowledge, fully-conjugate models incorporating both the distribution of temperature and flow in the jacket and simultaneously modelling the free-surface vortex in an unbaffled STR have not yet been investigated and this study aims to work towards this goal.…”