PTQ Q3 2023 Issue

column overhead temperature of 215°F. Detailed heater mod- elling and analysis determined that coil steam could be reduced from 29,000 lb/hr to 3,200 lb/hr without compromising heater reliability or run length. The large reduction in coil steam reduced the motive steam requirement from 122,000 lb/h to 83,000 lb/h. The large

Ejector system components

Components

Lb/hr

Mole weight Water vapour equivalent, lb/hr

% Total

Coil steam

3,200 9,960

18 18

3,093 9,627 8,720 2,250

13.1 40.6 36.8

Stripping steam Condensables Cracked gas/air

18,350

148

3,000

9.5

Total

34,500

23,690

100

Table 1

water side fouling, small changes in cooling water supply temperature or rate, or changes in condensable hydrocar- bon rate and composition. MDP margin allows higher con- densing pressure, higher first-stage inter-condenser outlet temperature and higher second-stage ejector process load and operating pressure before exceeding first-stage ejector MDP. MDP margin increases motive steam requirements. It can be tempting to minimise its allowance to reduce over- all total installed cost. History has repeatedly shown that designing with razor-thin margins in the ejector system can lead to millions of dollars in lost opportunity recurring over many years. The first-stage ejector process load consists of steam, condensable oil, and cracked gas/air. The vacuum heater coil steam injection rate, which is used to control oil res- idence time, accounted for more than 50% of the steam load. For deepcut operation, coil steam is often required to achieve reasonable heater run lengths. Its use allows higher heater outlet temperature before excessive tube coking. Coil steam, however, has only a small effect on flash zone oil It is essential during the troubleshooting effort to estimate an accurate first-stage process load from available ejector slop oil and non-condensable gas partial pressure because the heater outlet vapour and liq- uid phases separate in the transfer line and are not in equi- librium. Coil steam partial pressure has a small impact on increasing flash zone lift. Stripping steam rate, on the other hand, is an essential component in meeting yield because the trays intimately contact the rising steam and the flash zone liquid. For this reason, coil injection steam rate should be minimised to what is necessary for heater reliability. It should be calculated with robust methods, including tube- by-tube analysis, and not rules of thumb. Excessive use of coil steam directly contributes to larger than necessary first-stage ejector/intercondenser size and cost. The first-stage ejector process load has three compo - nents: process steam (heater coil, stripping steam, and sat- urated water in feed), condensable oil, and cracked gas/air leakage. For the revamp, cracked gas rate was set at 0.3 wt% feed and condensable oil rate was based on a vacuum

reduction in steam rate reduced the inter-condenser duty requirement. Two new exchangers sized with square pitch design could fit in the plot space occupied by the original exchangers. Table 1 shows the final ejector load that could be processed with a 19 mm Hg absolute suction pressure and 98 mmHg absolute discharge pressure (MDP). This design allowed for an 11 mmHg MDP margin which has proven to be sufficient for exchanger fouling, cooling water, and oil load deviations normally experienced in CV1. Occasionally, vacuum systems underperform because ejectors did not achieve design MDP. To avoid this scenario, ejectors should be properly tested. 4 In some cases, the ejec- tor steam nozzles can be moved forward in the steam chest to increase MDP at the expense of suction load capacity. In recent years there have been numerous failures resulting from the first-stage inter-condenser X-shell ‘long air baffle’ bypass. Bypass can occur from mechanical design issues or corrosion of the sealing mechanism.⁵ When the condensate temperature leaving the first-stage inter-condenser is much colder than the gas outlet, the ‘long air baffle’ is not work - ing and will likely lead to poor performance. 6 It is impor- tant when ejectors have a performance break to identify the root cause problem with complete field pressure and temperature data. It is essential during the troubleshooting effort to estimate an accurate first-stage process load from available ejector slop oil and non-condensable gas flow meters; otherwise, no reliable conclusions can be drawn. Vacuum heater modifications Restoring the vacuum column operating pressure to 19 mmHg absolute helped increase HVGO cutpoint. Yet, higher heater outlet temperature was needed to meet HVGO yield targets. Even though the absorbed duty was not increasing because the inlet temperature increased, the rate of coking would potentially increase for a higher out- let temperature. A thorough heater tube-by-tube analysis of film temperature and residence time was performed to determine what was required to ensure heater run length was not reduced. Vacuum heater tube coking depends on oil residence time and oil film temperature. Oil residence time is how long oil is in the heater tubes. Oil residence time depends on the tube size, oil per cent vaporisation, coil steam injec- tion rate, and location. An accurate vaporisation profile is needed to calculate the oil residence time, especially if it is analysed on a tube-by-tube basis. The vaporisation profile is highly dependent on the pressure profile. For this reason, the vacuum transfer line and heater tubes are modelled as

PTQ Q3 2023