School of Engineering


Master of Science in Engineering


Wayne Strasser


CFD, LDPE, PID-automation, Fuzzy Logic, Ethylene hot spots


Engineering | Mechanical Engineering


Computational Fluid Dynamics (CFD) was employed to develop a rigorous model of a low-density polyethylene (LDPE) autoclave reactor. Different numerical settings within the solver were evaluated to minimize false diffusion and to reflect the sensitive heat generation taking place during free radical polymerization. The rigorous CFD model employed reaction kinetics, Proportional Integral Derivative (PID) automated thermal management, and a rotating stirrer shaft. Validation was carried out to determine the sensitivity to time-step size, turbulence model, and grid resolution. Data were compared to an industrial scale plant autoclave to guide the development of CFD. Time-step independence was confirmed by comparing the moving time- and spatial-average temperatures across eleven thermocouples. The selected time-step size represents 1/130th of a stirrer revolution per time-step. A mesh refinement study revealed slight variation in the results between the base mesh of 6 million computational elements and the refined mesh consisting of 40 million. Ultimately, the variation between different grid resolutions was not significant enough to justify slowing down the solver speed by 14X by using the refined mesh. In a comparison of turbulence models, the shear stress transport (SST) model was found to predict higher concentrations of turbulent kinetic energy (TKE) resulting in a lower temperature distribution throughout the reactor than the differential Reynolds stress model (DRSM). The less diffusive DRSM was recommended for future studies. Increased rigor improved the model’s ability to match plant data, and CFD thermocouples were within 2.5% of temperatures from plant data. Next, CFD was used to study local decompositions in an LDPE autoclave reactor by identifying, characterizing, and tracking the trajectories 3 of contiguous hot spots (CHS). Local decomposition of ethylene occurs in very short time and spatial scales, potentially leading to thermal runaway and global decompositions as temperature and pressure increase beyond a recoverable threshold. CHS formed in the bulk flow but were quickly dissipating as they encountered strong mixing currents. CHS were found to have a high Biot number, describing that conduction was the limiting mechanism in removing heat. Overall, under stable operating conditions, the CFD model did not predict conditions which were conducive to global decomposition events resulting from CHS. Corroborating that result is the fact that current plant reactor employing the same operating conditions as CFD also does not report any recent decompositions. Finally, a new control technique utilizing a Fuzzy Logic PID (FLPID) controller was evaluated to determine its ability to shorten convergence times for the computationally expensive model. Catalyst feed rates were increased by 50% to evaluate the Fuzzy Logic controller’s performance during a process change. The nonlinear characteristics of the FLPID proved to get a model to quasi-steady state (QSS) 54% faster compared to the conventional PID controllers. Reducing the percent overshoot of the error and the rise time of the controller output, the FLPID demonstrated its ability to lower computational cost. The success of the FLPID in CFD offers the potential to improve control methods on actual plant scale autoclaves. In particular, the reduction in error overshoot could reduce the likelihood of reactor decompositions as the temperatures are kept within a tighter operational window.