Jun-2022

Importance of testing for vacuum ejectors in refinery service

Efforts should be made to identify and avoid errors in the specification, engineering, and manufacturing of vacuum system equipment before they manifest at start-up.

Edward Hartman and Tony Barletta, Process Consulting Services, Inc.
Laurent Solliec and Peter Trefzer, GEA Wiegand GmbH

Article Summary

Vacuum systems are critically important to the performance of refinery crude vacuum distillation units. Vacuum tower flash zone pressure is a result of vacuum system suction pressure plus pressure drop through the overhead vapour line and column internals. To maximise recovery of gasoil from vacuum residue, the flash zone should operate at the lowest possible pressure without exceeding tower capacity. Unfortunately, many vacuum towers operate above their design or expected pressure, resulting in lower vacuum gasoil yields and reduced profitability.

Many papers and articles in the technical literature discuss the performance and troubleshooting of vacuum ejector systems in refinery service.1,2. This article focuses on the importance of shop testing vacuum ejectors to ensure they meet their design parameters of pressure versus capacity and consequently prevent significant economic losses that can result from their underperformance. An additional benefit is the ejector performance curves derived from actual testing are paramount to any future troubleshooting, optimisation, and revamping of vacuum systems in operating units.

Crude unit vacuum systems
Multi-stage steam jet vacuum ejector systems are almost universally used to produce vacuum in refinery crude distillation units. They are particularly well suited to the large vapour volumetric rates present and high compression ratios normally required. The systems are arranged in two to four stages, with each stage consisting of an ejector and discharge surface condenser. Depending on unit capacity and desired flexibility, each stage may have ejectors or surface condensers in parallel. Steam works as the motive fluid providing the energy for compression. Load to each stage consists of non-condensable gas, condensable hydrocarbons vapour, and water vapour. Surface condensers minimise the quantity of condensable vapours and cool the non-condensable gas mixture flowing to the next ejector stage and leaving the system. This reduction in load results in smaller ejector size in the intermediate and final stages, as well as lower overall energy usage.

Discussed in further detail under vacuum ejector fundamentals, ejector suction pressure is a function of its load and performance curve, as long as its discharge pressure is below its maximum discharge pressure (MDP). In a multi-stage crude vacuum unit system, overall suction pressure is a function of vapour overhead load from the vacuum tower and first-stage ejector performance curve. This holds true provided the first-stage condenser, as well as ejectors, condensers, and interconnecting piping in subsequent stages do not cause the first-stage MDP to be exceeded.

In some cases, high vacuum system suction pressure, and consequently high tower pressure, is caused by process loads pushing the ejectors out on their curves. However, Process Consulting Services (PCS) has encountered several instances of ejector design errors leading to tower pressures up to 10 mmHgA above design right from start-up. When vacuum system fails to deliver from day one, system performance can further degrade rapidly with normal exchanger fouling and the initial 10 mmHgA miss can increase to 20 mmHgA or more above design suction pressure.

The impact of flash zone pressure on refinery profitability is magnified in units processing heavy crudes. Figure 1 shows vacuum residue yields for two different vacuum columns over a range of flash zone pressures. Unit 1 runs a blend of heavy Canadian crudes, while Unit 2 processes a blend of heavy Venezuelan and light US crudes. As pressure increases from 20 to 40 mmHgA, Unit 1 residue yield increases by about 3 LV% on crude and Unit 2 residue yield increases by about 4 LV% on crude. For a crude capacity of 100 MBpd, this corresponds to 3 and 4 MBpd of incremental vacuum residue, respectively. Considering a vacuum gasoil to residue downgrade penalty of $15/Bbl, the higher flash zone pressure results in profit losses of $16MM/yr and $21MM/yr.

Troubleshooting and diagnosis of crude vacuum systems underperformance can be difficult and time consuming. The process of determining ejector loads in an operating unit is full of uncertainties. It requires good field data, laboratory analysis, and flow measurements that are often unavailable. Furthermore, it relies heavily on process simulation models that are only as good as the simulation inputs. For units that have been in operation for more than a few months, additional uncertainties can further complicate efforts to narrow the root cause(s) of high vacuum column operating pressure. These include, but are not limited to, condenser fouling, ejector motive steam nozzle erosion, and tower internals damage that increases system load.

Therefore, every effort should be made to identify and avoid errors in the specification, engineering, and manufacturing of vacuum system equipment before they manifest themselves at start-up. In vacuum surface condensers, undersized shell nozzles, over-optimistic heat transfer coefficients, and vapour internal bypass are common problems. In vacuum ejectors, motive steam nozzle size and position are common culprits. In addition to generating accurate performance curves, shop testing vacuum ejectors provides a unique and valuable opportunity to prevent such problems.

Steam jet vacuum ejector fundamentals
Steam jet vacuum ejectors are essentially compressors with no moving parts. Figure 2 is a diagram of a steam jet vacuum ejector that establishes nomenclature and symbology used throughout this article.

The basic operating principle of a steam jet ejector is momentum transfer. The motive nozzle converts steam pressure energy into velocity energy, resulting in a supersonic velocity jet of steam that entrains the vapour and gas mixture from the suction chamber. The resulting mixture of motive steam and suction load enters the diffuser where velocity energy is converted back to pressure. At the diffuser throat, a pressure discontinuity or shock wave occurs which is responsible for an important property of steam jet ejectors — suction pressure is independent of discharge pressure up to a certain limit, and is only influenced by the amount of suction load. Figure 3 presents the performance curve of a large first-stage ejector, showing the relationship between suction pressure and load.

As mentioned above, suction pressure is independent of discharge pressure only up to a certain point. When discharge pressure becomes too high, the shock wave in the ejector throat can become unstable, leading to ‘broken’ ejector operation that produces high and sometimes fluctuating suction pressure. When an ejector is operating in a broken state, the suction pressure is unpredictable and depends on both suction load and discharge pressure. The discharge pressure that breaks the ejector is known as maximum discharge pressure, or MDP. MDP is commonly represented as a single number for a given ejector, but it actually varies with suction load. Figure 4 shows a combined plot of suction pressure and MDP curves versus suction load for an actual first-stage ejector. Note that the axes on this figure are reversed from the typical pressure versus load capacity curve, with X-axis indicating pressure and Y-axis representing load. This arrangement makes it easier to combine the individual capacity curves of a multi-stage system in the same plot.

Although ejectors can handle a wide variety of suction gasses with varying temperature and molecular weight, ejector curves and ejector testing are based on the principle of equivalent load. Any combination of suction vapours can be converted to equivalent water vapour (EWV) or dry air equivalent (DAE) load by simple corrections for molecular weight and temperature. Using equivalent load makes it easy to convert and plot any given suction load onto the equivalent load curve to check performance. Equivalent load can also be represented at different reference temperatures. This process is described in detail in industry standards.3,4 All loads discussed he