Effect of nozzle assembly, insulation and operating condition on partial combustion unit performance : computational fluid dynamics and experimental study

• Main Author: Law, Woon Phui
• Format: Thesis
• Language: English
• Published: 2017
• Subjects: TP Chemical technology
• Online Access: http://umpir.ump.edu.my/id/eprint/19534/19/Effect%20of%20nozzle%20assembly%2C%20insulation%20and%20operating%20condition%20on%20partial%20combustion%20unit%20performance%20%20computational%20fluid%20dynamics%20and%20experimental%20study.pdf; http://umpir.ump.edu.my/id/eprint/19534/

Partial combustion of syngas is often used to produce high-quality metallic iron in steel processing plant. Partial combustion unit (PCU) suffers from insufficient high temperature due to improper design configuration and heat loss. Consequently, higher rate of fuel injection and operating cost is required to maintain the quality of iron. Hence, a proper retrofit of design on the PCU is vital to offer better combustion performance and to reduce fuel consumption. This work aims to address the issues by proposing several new nozzle assembly designs and by installing thermal insulator to the existing PCU. Three new nozzle assembly designs were proposed to offer better turbulent mixing, whereas four different thermal insulators (i.e., brick, iron cast, ceramic fibre and graphite felt) were tested to minimize the heat loss. Experimental measurement is not applicable for opaque wall of PCU and hence computational fluid dynamics (CFD) technique is used to study the reactive fluid flow and heat transfer in the PCU. CFD simulation of the PCU at extreme temperatures is challenging owing to the coupling between the turbulence, chemistry and heat transfer. Hence, a CFD simulation must be validated before it can be routinely used. The validation was performed using the LDV measurement of a scale-down PCU rig, as well as by comparing with the published measurements. Three turbulence models, namely standard k-ε (SKE), Reynolds stress model (RSM) and scale-adaptive simulation (SAS) were employed. Combustion reaction was modelled using the eddy dissipation model (EDM) and detailed chemistry flamelet models, whereas the heat transfer was calculated by considering a convection, conduction and radiation. Discrete ordinates (DO) and spherical harmonics (P-1) were used for radiative heat transfer. The inlet of PCU and nozzle is treated as inlet velocity with species mixture while the wall is assumed as non-slip boundary. The finding showed that the non-premixed flame model with detailed chemistry provided a better prediction with 6.58% of deviation from measured temperature. It was found that the prediction without considering radiation yielded over 9% error compared to < 7% error when radiation is modelled. The DO model gave the best prediction of radiative heat transfer with 5.65% of deviation. Besides that, all turbulence models provided a good prediction of outlet temperature with a deviation of less than 6% from measured data. However, SAS gave the best prediction with 5.25% of error. All new nozzle assembly designs achieved over 11% higher temperature where the nozzle with flat surface wing gave the best performance with over 45% increase in temperature than the original wingless PCU. It was found that 40% increase in oxygen flowrate increased the peak temperature by about 12%. Dual-lance was found more effective than the single-lance operating at a similar oxygen flowrate. In addition, installation of insulation enhanced over 17% of peak outlet temperature. The best insulation was achieved using ceramic fibre with 20.4% higher peak temperature than the original non-insulated PCU. The finding from this work may useful for design retrofitting of a PCU, whereby the CFD model can be employed to simulate the PCU performance and the LDV data can be used for validation.