Search

Carbon steel (CS) was pre-treated in CO 2 -saturated methyldiethanolamine (aq) to form FeCO 3 surface films, the protectiveness of which against corrosion in monoethanolamine was studied using electrochemical impedance spectroscopy, open circuit potential measurements and ex situ surface characterisation. Only pure, dense FeCO 3 films were found to provide protection in lab-scale tests. Dissolved O 2 in pre-treatment solutions resulted in traces of iron (hydr)oxide within the films, exacerbating non-uniform corrosion. Impure films were compromised in pilot-scale experiments in the CO 2 capture plant at Imperial College London. Therefore, pre-treating CS with methyldiethanolamine is not a robust corrosion control strategy for industrial CO 2 capture plants. • Pilot-scale stability tests of protective films formed through pre-treatment were conducted. • Pure FeCO 3 formed by pre-treating carbon steel was protective against corrosion in ethanolamine. • Dissolved O 2 in methyldiethanolamine resulted in iron oxide within the film. • Iron (hydr)oxide in surface films enhanced non-uniform corrosion in ethanolamine. • Pre-treatment with methyldiethanolamine should not be used for corrosion control. [ABSTRACT FROM AUTHOR]

Abstract: Horizontal flows of two immiscible liquids with the same refractive index, a kerosene-like hydrocarbon and a glycerol–water solution, have been instigated with planar laser-induced fluorescence in a square duct. Four flow regime categories were observed, these being: (1) stratified flow; (2) mixed flow (i.e., two distinct continuous phase regions with droplets in each); (3) two-layer flow, comprised of a dispersed region and a continuous, unmixed region (i.e., oil-dispersed flow over glycerol solution flow and, oil flow over a glycerol solution dispersion); and (4) dispersed flows (i.e., continuous oil phase dispersion and continuous glycerol solution dispersion. The flow can be described as occupying three zones; an oil phase at the top, a glycerol–water phase at the bottom, with a mixed zone between them. The vertical height covered by the mixed zone increased for increasing superficial mixture velocity, and the vertical height of the glycerol–water phase decreased for increasing input oil fraction. At low oil fractions the interface level separating the two phases was not affected by changes to the superficial mixture velocity. However, at higher oil fractions the interface height from the bottom of the channel decreased progressively as the superficial velocity was increased. Higher velocities also gave rise to increasingly fluctuating interface level heights. The mean droplet size increased initially, reached a maximum and then decreased as the oil fraction was increased, and was largest at intermediate and smallest at high superficial velocities. [Copyright &y& Elsevier]

Abstract: A series of two-phase air–water experiments was carried out in order to study the initiation and the subsequent evolution of hydrodynamic slugs in a horizontal pipeline. Experiments were carried out at atmospheric pressure, 4.0bar(a) and 9.0bar(a), and the effects of superficial liquid and gas velocities were investigated. The test section used for these experiments is 37m in length, with an internal diameter of 0.078m. To study the interfacial development, measurements of interfacial structures were made at 14 axial locations along the test section, with data acquired at a sampling frequency of 500Hz. A large number of slugs were initiated within the first 3m of the test section, with the frequency subsequently reducing towards the fully developed value before the end of the pipe. This reduction in frequency was strongly influenced by the magnitude of the gas and liquid velocities. The frequency of slugging was not strongly affected when the system pressure was changed from 1 atmosphere, to 4.0 and 9.0bar(a), closely similar values being obtained at the 10 downstream locations. However, higher pressure delayed the onset of slug initiation, with “slug precursors” being formed further downstream as the pressure was increased. The statistical distributions of slug lengths and of the time intervals between slug arrivals were examined in detail and compared to several standard distributions. This showed that slug initiation may be reasonably approximated as an uncorrelated Poisson process with an exponential distribution of arrival times. However, once slugs have developed, there is strong correlation and the arrival time intervals, as well as the lengths, are best represented by the log-normal distribution. [Copyright &y& Elsevier]