Feb-2023

De-mystifying vacuum ejector systems

First-stage inter-condenser design and performance are critical for reliable vacuum system performance.

Scott Golden, Tony Barletta and Steve White
Process Consulting Services, Inc.

Article Summary

Vacuum systems are critically important to maintain column operating pressure and maximise distillate product yield and economic potential. Yet systems continue to operate poorly even upon initial installation or after turnaround modifications. Often, performance break increases column operating pressure by 20-40 mmHg, dramatically reducing distillate yield. Performance break occurs because the first-stage ejector operating discharge pressure is at or above its maximum discharge pressure (MDP). There are many potential causes of poor vacuum ejector system performance, yet an increasingly common one is failure of the first-stage inter-condenser to operate as intended.

This article will highlight the importance of the first-stage inter-condenser and demonstrate how failure to perform as intended, specifically the sub-cooling zone, frequently results in performance break and unstable, high column operating pressure. Fundamental concepts including inter-condenser design, ejector performance curves, MDP, and interaction between system components are reviewed.

The critical importance of ejector system design-phase considerations, including cracked gas production, first-stage ejector MDP margin, and system pressure drop, are reviewed. System components interact; first-stage ejector suction pressure is directly tied to the second-stage ejector suction pressure (see Figure 1). There are many potential causes of first-stage inter-condenser poor performance; an increasingly common one and the focus of this article is failure of the first-stage inter-condenser to sub-cool.

Ejector system design optimisation
Most refinery vacuum unit ejector systems are three- or four-stage units. A stage is defined as an ejector and condenser in series. Vacuum ejectors use motive steam energy to raise the process load from the column top pressure to the outlet gas discharge pressure. The condenser for each ejector stage condenses steam and oil to reduce the load to the downstream ejector. The cooling water (CW) rate is set by the first-stage inter-condenser requirements. The design of the vacuum system is typically optimised to minimise motive steam and CW consumption. The first-stage ejector consumes up to 70% of the total motive steam. Consequently, the first-stage ejector and inter-condenser are much bigger in size than subsequent stages. They typically represent more than 50% of the total installed cost (TIC) of a typical vacuum system. Multistage vacuum system design focuses on first-stage design conditions because the first stage costs more and consumes more steam and CW than the rest of the system. For that reason, they represent the biggest opportunity for optimisation and savings.

An ejector uses motive steam at a much higher pressure than the suction pressure to entrain the process stream into the steam chest (see Figure 2), where it is boosted from suction pressure Ps to discharge Pd at the ejector outlet. Ejector operating discharge pressure Pd is set by the downstream ejector process load and system pressure drop. An ejector’s design and performance can be related by a set of key parameter ratios, compression, entrainment, and expansion, as defined in Figure 2. These ratios are used to specify the required size, shape, and motive steam requirement.

For a given suction load and pressure, two parameters that impact first-stage ejector sizing are motive steam pressure and MDP. Higher pressure motive steam will generally require less steam than lower pressure steam. However, in most refineries, steam header pressure levels are set, and the ability to optimise is limited. In rare cases, the selection of steam conditions can be used to optimise ejector design.

Once the motive steam pressure is set, the first-stage ejector discharge pressure is optimised to minimise the size of the ejector and the amount of motive steam it will use. This relationship is reflected in the compression ratio Pd/Ps, where Pd is MDP and Ps is the suction pressure. As the compression ratio increases, so does the required motive steam. Designing for higher MDP increases motive steam consumption. With end users not providing adequate or any design margin guidance, vacuum system suppliers design the system with razor-thin margins to reduce steam and CW consumption. They do this in part because selection criteria will often include utility consumption comparisons. In many cases, the MDP is so low that even small changes in design conditions or underperformance of the first-stage inter-condenser will lead to performance break. Even worse, when the first-stage inter-condenser fails to sub-cool there is a guarantee that the system will experience performance break.

First-stage inter-condenser design evolution
Modern large first-stage inter-condensers are supplied as X-type shells that include a sub-cooling zone. This design is an evolution from older and smaller designs that used E-shells. The advantage of the X-shell is that it decreases pressure drop compared to E-shell, lowering steam consumption. The X-type first-stage inter-condenser is designed to sub-cool the outlet vapour below the bulk outlet condensate (see Figure 3).1 This is accomplished with the ‘long air baffle’. The long air baffle’s function is to force the outlet vapour across exchanger tubes with the coldest CW, thereby sub-cooling the gas below its condensate. Roughly 20-25% of the exchanger tubes are located under the long air baffle and are dedicated to sub-cooling.

The sub-cooling zone is critical. When working as intended, the sub-cooled zone minimises second-stage ejector gas load. The selection of the ejector MDP is based on the load to the second-stage ejector commensurate with the sub-cooled zone working as intended. In practice, when the first-stage inter-condenser does not sub-cool the outlet vapour, the outlet vapour to the second-stage ejector increases above its design point. The higher load increases the suction pressure of the second-stage ejector, which ultimately leads to higher first-stage ejector discharge pressure. When the first-stage ejector discharge pressure exceeds its MDP, performance break will occur. Mechanical design flaws that fail to seal the long air baffle or damage, corrosion, or improper installation of the seal strips are the root cause of the inter-condenser’s inability to sub-cool outlet vapour and overload the second-stage ejector. A performance break is guaranteed if the ejector system was designed with minimal or inadequate pressure drop margin. Therefore, the first-stage inter-condenser must perform as intended and sub-cool outlet vapour to avoid performance break.

Easy to diagnose, hard to troubleshoot
Long air baffle bypass is easy to diagnose because the vapour leaving the first-stage inter-condenser should be colder than the condensate draining from the bottom of the shell. When the temperatures are inverted, that is an indication of a bypass around the sub-cooling zone (see Figure 4). Troubleshooting vacuum systems is inherently difficult because much of the information required is not readily available. It is more the norm than the exception for vacuum systems to be poorly instrumented. Flow meters for key streams are often missing. Conclusions made with incomplete data or data based on design first-stage gas loads or using original vacuum system design information 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 condensate because the difference may be only 3-4°F.