stage test measuring points and its respective translation to the design conditions is exemplified in Table 3 . All ejector elements in this sys- tem were tested by one of the methods illustrated above. This acceptance testing allowed GEA to demonstrate the expected design performance of each ejector, as well as to generate accurate suction and MDP performance curves. Overall vacuum system performance and its impact on post-revamp unit operations will be discussed in a future article. Conclusion Acceptance or shop testing of vac- uum ejectors is used to generate an accurate set of ejector perfor- mance curves. More importantly, it provides an opportunity to pre- vent ejector underperformance by identifying and correcting design or manufacturing errors. Given the significant profit loss that can result from a ‘broken’ ejector in a crude vacuum distillation unit, shop test- ing of new ejectors should be man- datory. This is not a topic to cut corners on cost. References 1 Lieberman N, Cardoso R, Troubleshoot operation of a steam ejector vacuum system, Hydrocarbon Processing, Feb 2016, 59-64. 2 Cantley G, et al , Maximise VGO yield, PTQ Revamps & Operations , 2005, 22-25. 3 Heat Exchange Institute, Inc., Standards for Steam Jet Vacuum Systems. 4 German Standard DIN 28 430, Rules for the Measurement of Steam Jet Vacuum Pumps and Steam Jet Compressors. Edward Hartman is a process engineer with Process Consulting Services Inc, in Houston, TX, USA. He has 35 years of experience in studies and design packages of refinery units, specialising in distillation equipment design. Tony Barletta is Vice President with Process Consulting Services. Inc. He has over 34 years of experience with refinery revamps and process design for heavy oil units. Laurent Solliec is an R&D Manager at GEA Wiegand GmbH, Germany. He specialises in customer oriented solutions and ejector technology, and heads the test bench facility. He holds a PhD in fluid dynamics engineering. Peter Trefzer is Product Manager for vacuum systems and ejectors at GEA Wiegand. He specialises in multi-stage ejector applications and holds a degree in mechanical engineering.
1st stage model test results and translation to full-size ejector
Model test results
Full-size ejector results
m· s,test lb/hr
P m,test psia 194.4
P s,test
P d,test
m· s
P m
P s
P d
mmHgA
mmHgA
lb/hr 9,634
psia
mmHgA
mmHgA
89
11.1
90.0
194.4
11.1
90.0
Table 2
3rd stage ejector scaled test results and translation to design conditions
Scaled test results
Design test results
m· s,test lb/hr
P m,test psia
P s,test
P d,test
m· s
P m
P s
P d
mmHgA
mmHgA
lb/hr 1,408
psia 128.8
mmHgA
mmHgA
325
28.7
27.8
123.8
111.0
495.0
Table 3
compiled over decades, to deter- mine this empirical factor. Design motive steam conditions include superheat, which can- not be delivered on the test stand. Therefore, the model ejector is fit - ted with a motive nozzle throat area sized to account for the higher den- sity of the saturated steam used on the test stand, such that the correct motive steam mass rate is deliv- ered. Test results from one of the measuring points and its respective translation to the full-sized ejector is exemplified in Table 2 . 3rd stage ejectors – scale test with adjustment for steamsuperheat The third-stage ejectors physically fit in the test stand but their motive steam rate exceeds test stand capacity. Therefore, a scale test is performed with a scale factor of 4 to 1 ratio. Again, the design motive steam conditions include super- heat, which cannot be delivered on the test stand. In order to achieve the same motive steam mass rate through the motive nozzle, a super- heat temperature correction factor calculated from steam properties is applied to the motive pressure in conjunction with the scale factor. As previously mentioned, adjust- ing steam pressure to account for superheat results in a different expansion ratio Er between design and test conditions, while the com- pression ratio Cr remains constant. Therefore, the entrainment ratio ω is adjusted by an empirical fac - tor based on GEA’s experience. Test results from one of the third-
Crude vacuum unit ejectors testing A revamp designed by PCS of a major US Gulf Coast crude vacuum unit illustrates the application of the testing methods. Unit charge con- sists of atmospheric residue from a blend of heavy Venezuelan and light US crudes. In order to max- imise refinery profitability, one of the main revamp objectives is to improve heavy vacuum gasoil (HVGO) cutpoint to over 1050°F, thus minimising vacuum residue yield. Therefore, required vacuum unit design conditions are stringent: heater outlet temperature in the 780°F range, stripping steam rate in excess of 8.0 lb/bl of residue and 30 mmHgA vacuum tower flash zone pressure. The four-stage existing vacuum system was revamped with all new ejectors, as well as new first- and second-stage intercondensers. 1st stage ejectors –model test with adjustment for steamsuperheat The first-stage ejectors cannot be directly tested due to their sheer size, requiring a model test. The model ejector is scaled down based on the similitude laws, resulting in a scale factor of 100 to 1 ratio. As previously mentioned, the model ejector is less efficient than the actual ejector due to wall friction effects. This efficiency difference is accounted for on the translation of each test point to the full-sized ejector curves by an empirical fac- tor. Although the friction correc- tion factor is correlated to Reynolds number, in practice GEA relies on an extensive database of test results,
46 PTQQ 2 2022
www.digitalrefining.com
Powered by FlippingBook