API PUBL 932-A:2002 pdf download

API PUBL 932-A:2002 pdf download.A Study of Corrosion in Hydroprocess Reactor Effluent Air Cooler Systems.
The ability of both anunomum chloride and ammonium hisulfide to condense as solids from the vapor phase and thereby cause blockage of the flow path was the motivation to introduce a water wash to soluhilize the deposited material. Unfortunately, the resulting aqueous solutions are extremely corrosive unless substantially diluted. and are in fact the cause of the corrosion problems in these systems.
The survey gathered data on the chemical composition of the effluent stream including contaminants, and attempted to define the corrosivity of the aqueous phase by factoring in the amount of water added to the system and its velocity. The concentration of bisullide solution was measured in most cases at the downstream water separator and this value was used as a measure of the aggressiveness of the process stream. It was observed thai at a concentration of 2% bisultIde or below corrosion was mild but at 3% – 4% or more, significant corrosion began to occur.
The results helped to SUpport a previously suggested relationship between the bisulfide concentration and velocity. wherein the hisulfide level was represented by the product of mole% NH x mole% HiS. designated as the K1, (Pichl) factor. It was proposed that for K,, values of 0.1 —0.5, velocities in the range 15 Ills — 20 ft!s would be appropriate hut for K,, values above 0.5 there was no suitable velocity. The higher the K,., value the tighter the tolerance on velocity. It was also found that velocities of 10 Ills — 12 ft/s could result in stagnant deposits underneath which severe corrosion could occur, hence a lower limit of 10 Ills was suggested. with an upper limit of 20 ftls and an optimum of 15 Ills.
A major conclusion drawn from this survey was thai air cooler corrosion is a complex phenomenon having numerous interdependent variables. This reduces the prospects of successfully eliminating corrosion by control of one or even several of the variables.
4 1996 UOP Survey
In the 20 years following the NACE survey the problems with corrosion continued, giving risc to a number of publications discussing various aspects of the phenomenon3’. Air cooler tubes continued to be the principal focus of the discussions although piping problems were also recognized. Laboratory studies of corrosion were unable to clarify the use of parametric variables in controlling corrosion6. Thus, in 1996. UnocalIUOP initiated a survey of its licensees to expand the experience data base and possibly identify any new betors in the corrosion of REACS and related piping.
The survey consisted of a comprehensive 10-page questIonnaire covering a variety of process and mechanical design information and corrosion cxpcrience. Topics included general operating conditions such as process mode. feedsiocks, gas flow rates, contaminants (l-12S. NH3. Cl. CN). water wash details, and flow velocities. Air cooler and piping design and layout. materials and corrosion experience were also covered.
Forty-six responses were received from operators of five different types of hydroprocess unit. The inftwrnation in the responses was compiled into tabulated fomiats and analyses made of the effects of certain variables on the com)sion experienced. Not all the respondents were able to provide values for each of the key parameters requested so that UOP had to provide estimated values by calculation from key operating data such as feed quality, charge rates, reactor efficiencies, flow rates, temperatures. pressures. and tube and piping flow areas. Note that these calculations, in particular velocities, were not based on rigorous process simulation but rather on factored estimates based on representative models for cich unit configuration. To ensure consislencv, UOP calculated values for K,,, NH4HS concentration and velocities tbr all of the units.
The results were presented in Paper 490 at Corrosion 977, The diversity in the responses is illustrated in Table 1, which summarizes the range of values received br the key sariables: however, it cannot be construed that such a range of values in the key parameters will necessarily result in wide pattern of corrosion behaviors because of the interdependency of corrosion on several parameters at the same time. Only if all the parameters are at one end of the range or the other can extreme hehas ior be anticipated In addressing corrosion of carbon steel air cooler tubes, the effect of K, factor on the severity of corrosion was evaluated by plotting K,, factor versus level of corrosion experienced. The level of corrosion was associated with tube life and the following rankings used.