PTQ Q1 2023 Issue

A look at the condensing curve helps understand what is happening in the inter-condenser. The first-stage inter- condenser design duty was 65 MM Btu/hr. A large portion of the steam and condensable oil condenses in the first- stage inter-condenser. The non-condensable hydrocarbon does not condense and exits as vapour, along with some steam and condensable oil. The condensing curve (see Figure 11 ) shows a desuperheating zone on the inlet and a sub-cooling zone on the outlet. The majority of the duty is for steam condensing and occurs over a relatively con- stant temperature. Only the last 5% of exchanger duty occurs under the long air baffle, where the temperature changes by as much as 12°F due to increasing cracked gas/ air partial pressure. While this duty is relatively small, its effect on second-stage ejector load is high because 5% of the exchanger duty in this case represents over 3,000 lb/ hr steam that is condensed in the sub-cooling zone. The design second-stage process load is 8,000 lb/hr steam equivalent. A relatively large percentage of the tube bundle area is dedicated to sub-cooling. It is, therefore, imperative that the long air baffle seals as intended. A large portion of the heat transfer and condensation occurs with a very low driving force, given that the CW temperature varies from 90°F to an outlet of 103°F. The ini- tial steam condensation temperature is 117°F. This is why ejector systems can be very sensitive to CW conditions. If the water flow rate is below design or the temperature above design, condensation may suffer, leading to higher second-stage ejector loads. When this occurs, the pres - sure increases and the bulk condensation temperature will increase by a few degrees. Nonetheless, the outlet vapour temperature will be less than the condensate if the sub- cooled zone is not bypassing. This highlights the impor- tance of accurate field data when troubleshooting vacuum systems. General fouling and long air baffle bypass are not the same! General fouling and heat transfer impacts The first-stage inter-condenser can be subject to tube and shell-side fouling and other items such as higher-than- expected condensable oil quantities or even liquid entrain - ment from the top distributor. All of these items have the same effect – they reduce the overall heat transfer coefficient. When one or more of these occur, and depending on the severity, the flow rate of vapour leaving the inter-condenser increases, coupled with an increase in operating pressure and temperature. The condensate and vapour outlet tem- perature will both increase but there will still be sub-cooling. The vapour temperature will be below the condensate as long as the long air baffle is working as intended. It is not uncommon for crude column stripping trays to get dislodged during start-up or during the run. A loss of bottoms stripping leads to an increase in condensable vapour leaving the top of the vacuum column. Should the vacuum system be designed to overcome this? The answer is yes. Appropriate design margins including inter-con- denser fouling factors, cracked gas flow rate, and, most importantly, MDP margin help avoid performance break for routine things that happen in crude units.

125

120

115

110

105

100

0 1020304050607080 Percent duty 90

90 100

94 mmHg 0.1 wt% NC 94 mmHg 0.3 wt% NC

84 mmHg 0.1 wt% NC 84 mmHg 0.3 wt% NC

Cooling water

Figure 11 First-stage inter-condenser condensing curve

the gas outlet temperature to have an 8-10°F approach to the CW inlet temperature. This is accomplished with the sub-cooling zone utilising the long air baffle. Field measure - ments for many operating vacuum systems have shown a 2-4°F gas outlet sub-cooling compared to the condensate outlet when the long air baffle works. When it does not work, Figure 4 shows that gas outlet temperatures can be as much as 20°F higher than the CW inlet, dramatically increasing second-stage gas load. The second-stage ejector design capacity is based on the first-stage condenser outlet vapour rate. For this par - ticular vacuum system, the outlet vapour rate is 8,000 lb/h. Roughly 3,000 lb/h of additional steam is condensed in the sub-cooling zone. Without it, the load to the second-stage A relatively large percentage of the tube bundle area is dedicated to sub-cooling. It is, therefore, imperative that the long air baffle seals as intended ejector would be 11,000 lb/h. The sub-cooling zone obvi - ously does a very good job of reducing the second-stage ejector load. However, the long air baffle adds mechanical complexity because it must be sealed to prevent any bypassing of the sub-cooling zone. A small mechanical defect or seal strip leak which leads to a long air bypass can easily result in performance break, especially if the MDP design margin is low. For example, if the MDP margin for this first-stage ejector was 3 mmHg, it would take only 300 lb/h of long air baffle bypass to increase the second-stage ejector suction pressure by 3 mmHg and cause performance break of the first stage (see Figure 8). It is obviously critical that the sub- cooling zone performs as intended and that care is given to making sure the long air baffle is fully sealed.

PTQ Q1 2023