rarely identify root cause problems. More often than not, these approaches reach absurd conclusions. Troubleshooting requires accurate field measurements of pressure and temperature (see Figure 5 ). High-precision vacuum pressure gauges are required to measure small pressure drops. For example, the design first-stage inter- condenser pressure drop will typically be 3-6 mm Hg. Care must be taken to measure temperature. It is not uncommon for standard K-type thermocouples to be off a few degrees of each other at the temperature ranges encountered. This is important when comparing outlet vapour to the conden- sate because the difference may be only 3-4°F. Would you evaluate pump performance without knowing the flow rate, pressure drop, and comparing the measured operation to the certified pump curve? When a vacuum system is not performing per design, a comparison of the first-stage ejector load and pressure profile is vital. For this reason, it is helpful to have flow meters on the motive steam, off-gas, as well as the hot-well oil and water streams. Sample locations must be available to analyse composition. And finally, certified ejector performance curves must be available. Without real data, it is impossible to evaluate the system shown in Figure 5. Criteria for first-stage ejector suction loads Suction loads to the first-stage ejector must be established when designing a new or revamping an existing vacuum system. The process load to the ejector is the vapour leav - ing the top of the vacuum column. The outlet vapour is divided into the following components: steam, condens- able hydrocarbon, and cracked gas/air leakage. The flow rate and molecular weight of each are provided to the ejec - tor manufacturer. The ejector manufacturer will convert the flow rate to air or water vapour equivalent using Heat Exchange Institute (HEI) methods or, in some cases, DIN standards. The process loads are converted to equivalents to allow for direct comparison to ejector test curves since the testing process does not use hydrocarbon vapours. Table 1 shows first-stage ejector suction loads for a large vacuum system. The corresponding first-stage ejector curves are shown in Figure 7 . The Y-axis process load is based on the water vapour equivalent. In this case, the pro- cess steam and condensable oil rates represent 90% of the first-stage process load and largely determine the motive steam and CW requirements. Process steam can comprise bottoms stripping steam, heater velocity steam, and saturated water in the crude column bottoms product. The variables of stripping steam, heater outlet temperature, and column operating pres-
Vacuum column overhead
1st stage ejector
1st stage inter-condenser
Pressure, mmHg Temperature , o F
2nd stage ejector
Condensate
Figure 5 Ejector system field data
or coil steam have less condensable hydrocarbon in the overhead vapour. The second component is cracked gas/air leakage. Cracked gas (or non-condensable gas) is generated in the vacuum heater tubes. A small allowance is included for air leakage into the column. The cracked gas flow rate is much higher than air leakage. Hence, the two are generally lumped into one term. Cracked gas is generated in the vacuum heater tubes and is a function of heater design, burner operation, heater outlet temperature, and crude type. Estimating the cracked gas flow rate precisely is very difficult. There are too many variables to estimate the cracked gas flow rate accurately. For example, a cracked gas rate measured on an existing unit with the same crude blend may be used but must be adjusted for differences in heater outlet tempera - ture and design. Heater design plays a significant role in cracked gas production. A well-designed heater with mini- mum oil residence time and high oil mass velocity will have lower film temperatures and subsequently lower cracked gas rate than a poorly designed heater. Often, the cracked gas rate is, at best, an educated guess. The cracked gas rate is normally a small percentage of the total suction load to the first-stage ejector and has a relatively small impact on the size of the first-stage ejector and inter-condenser; therefore, it is prudent to apply a generous margin. The condensable oil rate is a function of vacuum tower top pressure and temperature, amount of steam used, and atmospheric crude column bottoms stripping per- formance. Vacuum column overhead temperature must
sure are optimised to achieve HVGO cut-point objectives. Of the three ejector load components, process steam is the easiest to quantify because they are well defined during the vacuum unit pro - cess design phase. When vacuum heater velocity steam and vacuum column stripping steam are used, the condensable oil rate leaving the top of the vacuum column will be higher because of the steam partial pressure effect at the top of the col- umn. Dry vacuum units that do not use stripping
First-stage ejector components
Components
lb/hr Molecular,
Water vapour equivalent, lb/hr
% Total load
wt 18 32
Process steam Cracked gas/air Condensable oil
13,160
12,720
53.7
3,000
2,250 8,720
9.5
18,350 34,510
148
36.8
Total
23,690
100.0
Table 1
31
PTQ Q1 2023
www.digitalrefining.com
Powered by FlippingBook