Gas 2025 Issue

Gas from 1st 1/2 of area Gas from 2nd 1/2 of area Gas from absorber

10 , 000

Gas from 1/4 of area Gas from 3/4 of area Gas from absorber

1 , 000

1000

100

L/G in 1/4 of area (gpm/MMscfd)

L/G in 1st 1/2 of area (gpm/MMscfd)

tower’s entire cross-section. However, a deviant solvent flow to one-quarter of the absorber (with the rest to the other three-quarters) affects the two sections to an une - qual extent. H2S treating is not a linear function of compo - sitions and flow rates. As Figure 4 shows, H2S treatment from the one-quarter-area section is much more negatively affected by reduced L/G ratio (note the logarithmic scale on the y-axis of these plots). However, Figure 5 shows that when one of the phases is maldistributed in fully one-half of the absorber cross- section, the two columns in the parallel column model (and the two sections of the real column) become indistinguishable. Perfect phase distribution is optimum, although performance loss is more severe when either phase is maldistributed over half the absorber cross-section. In either case, though, in this example, only a relatively small degree of maldistribution pushes treating to exceed 100 ppmv H2S. In other words, treat - ing performance is quite sensitive to phase maldistribution. Hidden causes Qualitatively, the foregoing offers the possibility of some guid - ance as to what effect maldistribution might have on TGTU performance, but quantitatively it is predicated on being able to assess actual flow rates to various areas across the column cross-section, something we are unable to do. However, if the tower’s actual mass transfer performance is known, this type of analysis offers the possibility of estimating how much maldistribution might be required to produce it. However, which of the gas or liquid is mechanically maldistributed and which is a consequence of the other will need additional information (for example, measured dis - tributor out-of-levelness or perhaps even a computational fluid dynamics [CFD] study of the gas inlet flow in the tower sump). Two parameters are needed: the relative cross-sec - tional areas affected and the deviations of the phase flows from their perfectly distributed values. There is no obvi - ous way to determine the latter, but relative areas might, in principle, at least, be roughly estimated using thermal images taken from different orientations. Figure 4 How solvent and gas maldistribution expressed as L/G ratio (gpm/MMscfd) in the ¼-area column affects H2 S content of the treated gas from the ¼- and ¾-area parallel columns and from these columns when combined into the real absorber

Unfortunately, TGTUs operating selectively do not usu - ally show the large temperature changes exhibited by CO2 absorbers, for example,5 probably making thermal imaging unworkable except in the rare instance of a TGTU pro - cessing a high-H2S, low-CO2 gas or when maldistribution is extreme. Nonetheless, maldistribution can be a hidden cause of severe failure to properly treat tail gas and meet SO2 emission regulations, and the parallel column model offers a way to analyse it. References 1 Weiland, R. H., Hydraulic stability of dualflow trays , paper presented at AIChE Spring National Meeting, New Frontiers in High-Capacity Tray Technology, Houston, TX, 2001, p12. 2 Sun, C. G., Yin, F. H., Afacan, A., Nandakumar, K., Chuang, K. T., Modelling and simulation of flow maldistribution in random packed columns with gas-liquid countercurrent flow, Chemical Engineering Research and Design, 78 (3), 2000, pp.378-388. 3 Kooijman, H. A., Zhou, J., Taylor, R., Application of a new parallel col - umn model to simulating maldistribution in packed columns, Chemical Engineering & Processing: Process Intensification , 171, 2022, 108436. 4 Schultes, M., Influence of liquid redistributors on the mass trans - fer efficiency of packed columns, Ind. Eng. Chem. Res., 39, 2000, pp.1381-1,389. 5 Cooper, E., Weiland, R., Reducing CO2 slip from the syngas unit of an ammonia plant, paper presented at Nitrogen and Syngas 2016, Berlin, Germany, February 29-March 3. Simon Weiland is Product Manager responsible for the development and marketing strategies for ProTreat, SulphurPro, and ProBot. He holds a BS in chemical engineering from the University of Oklahoma and is an expert in the fundamentals-based modelling of gas treating and sulphur recovery operations. Email: simon.weiland@OGTRT.com Prashanth Chandran is Technical Development Team Lead. He holds a BTech in chemical engineering from Anna University, India, and an MS in chemical engineering from Oklahoma State University. Ralph Weiland is OGT’s Chairman, having stepped down as CEO in June 2021. He led the development of the ProTreat process simulation package. He holds BASc, MASc, and PhD degrees in chemical engi - neering from the University of Toronto. He taught chemical engineer - ing for 30 years at universities in Canada, Australia, and the US and directed graduate research in gas treating. Figure 5 How solvent and gas maldistribution expressed as L/G ratio (gpm/MMscfd) in the ½-area column affects H2 S content of the treated gas from each parallel columns and from these columns when combined into the real absorber

Gas 2025