PTQ Q2 2022 Issue

Cr, Er, and ω is valid, regardless of the absolute values that make up those ratios and with the proper correction factors. indus- try standards cover many testing related topics in detail. The stand- ards are written to cover testing of ejectors for a wide range of ser- vices. Ejectors designed for refinery vacuum systems normally bring additional challenges due to size and utilities consumption. Their testing should include a range of suction loads, with suction pres- sure and MDP recorded at each load point. Although formal test- ing guidelines allow testing only three points, adding additional test points is trivial after going through the effort of setting up the ejector on the test stand. Depending on unit process design and modes of operation, it may also be advisable to check unusual loads correspond- ing to expected process swings. Ejectors that may operate during periods of very low load should be tested all the way down to zero load to ensure suction pressure stability. In addition, the zero-load suction and discharge pressures are essential at start-up of the vacuum column for comparison with in-situ measurements. Test stand setup and limitations Previously referenced Refinery ejectors are typically tested with steam as load rather than air. Figure 5 shows a represent- ative test stand setup. This test setup allows for control of motive steam pressure, suction load, and ejector discharge pressure. Motive steam is let down to the test pressure through a pressure control valve. If the vacuum system design condi- tions specify superheated motive steam, then superheat may be added by an electric superheater or motive steam pressure can be adjusted to account for the lack of superheat. 4 Suction load is metered by adjust- ing steam pressure upstream of a Heat Exchange Institute (HEI) noz - zle under critical flow. Ejector dis - charge is routed to a direct contact condenser where motive and load steam are condensed by mixing with cooling water. Condenser pressure is maintained by a downstream

10000 11000 12000 13000 14000 15000

0 3000 2000 1000 4000 5000 6000 7000 8000 9000

Suction

MDP

0

10

20

30

40

50

60

70

80

90

100

110

120

Pressure (mmHg absolute)

Figure 4 Suction pressure and MDP vs load performance curves

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 molecu- lar 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 herewith are on an EWV basis. Ejector similitude and scaling Ejectors are commonly defined by the following three non-dimen- sional parameters:

Expansion ratio: Er = P m P s Entrainment ratio: ω = m . m m . s

Ejectors are scalable devices in terms of both physical size and non-dimensional parameters. Similitude laws allow ejectors to be scaled by applying appropri- ate ratios to all critical geomet- ric parameters, including throat area, motive steam nozzle throat area, and motive steam nozzle dis- tance to throat, among others. A family of ejectors sharing identi- cal geometric ratios also shares all combinations of Cr, Er, and ω. The entrainment ratio ω is not exactly constant because it is affected by friction through the throat – an ejec- tor with a larger throat always per- forms better than a similar scaled down ejector with a smaller throat. Therefore, for a defined ejector geometry, testing at any set of fixed

Compression ratio: Cr = P d P S

Liquid ring vacuum pump or atmosphere

Cooling water supply Air

PI

TI

PI

Motive steam

Superheater (optional)

Direct contact condenser

Separator

TI

PI

Load steam

Cooling water return

Separator

Condensate drain

HEI nozzle

Condensate drain

Figure 5 Test stand configuration

44 PTQQ 2 2022

www.digitalrefining.com

Powered by