Page 238 - Subyek Teknik Mesin - Forsthoffers Best Practice Handbook for Rotating Machinery by William E Forsthoffer

Be st Practice 3 .26 Compressor Best Practices
B.P. 3.25. Supporting Material subjected to flow conditions. It can be argued that if a perfor-
mance test is conducted in accordance with ASME PTC-10, the
compressor will be tested under some load, which will serve to
Mechanical test requirements prescribed by API 617 for cen- eliminate the concerns noted above. However, since API 617
trifugal compressors can be met under full load, limited load or does not set limits on the performance test equivalent speed,
no load (vacuum). No load testing (vacuum testing) can be ac- a performance test is usually run at speeds far from design (50%
complished with lowest set up time and operating cost. Given e 65% of rated speed). Lower operating speeds do not duplicate
the recent amount of compressor shop load, vacuum mechanical field aerodynamic forces on the rotor of centrifugal forces.
testing has become the choice of compressor suppliers. Since To ensure optimum field reliability, users should always re-
API 617 does not outlaw shop mechanical vacuum tests, end- quire that a mechanical shop test be run under a minimum of
users are left to either accept these tests (sometimes as a sur- 10% load and that the shop-testing program will subject the
prise) or to take a proactive approach in the job specifications to rotor to operating temperatures. It has been our experience that
prohibit vacuum mechanical tests. the most costly revenue-reducing field problems are related to
If mechanical vacuum testing has been performed for years rotor thermal expansion issues and aerodynamic instabilities.
(since the 1960s), then why is it a concern? Today, with mega Thanks to the extensive rotor-dynamic requirements of API
process units and the high cost of products, daily plant pro- 617, field rotor response issues have been significantly reduced in
duction revenues have never been higher. A day’s production recent years and are not of concern if all rotor system design issues
loss can exceed millions of dollars. Shop testing therefore must are met. An argument can be made for only mechanically testing at
closely mirror the field conditions that the compressor will ex- 10% load to ensure aerodynamic forces will not be present during
perience, to ensure that the installed unit will achieve the field operation. Considering the present amount of industry ex-
highest level of reliability. perience with high-density compressor applications (reinjection,
Shop mechanical tests conducted under vacuum do not du- recycle and synthesis gas), we offer the following guidelines for
plicate any field conditions other than shaft speed. Impellers do determining if a full pressure mechanical test should be performed.
not experience torque, temperature or thrust loads. Rotor as- A full-pressure shop mechanical test is required if:
semblies are not subjected to thermal expansion. Journal bear-
ings are not subjected to aerodynamic forces. Thrust bearings The vendor does not have experience with this pressure level
are not subjected to axial loads. Seals are not subjected to op- at similar gas densities.
erating sealing pressures. If experience at similar densities is not with the gas path
In addition, many centrifugal compressor vendors are using components to be used for the project (vaned diffusers and
vaned or low-solidity diffuser vanes to increase head produced aerodynamic stability devices).
and compressor efficiency. While these components usually One final note regarding full-pressure tests: these tests,
achieve their objectives, they can produce aerodynamic in- designed to demonstrate that the compressor will be free of
stability, which can result in high sub-synchronous vibration aerodynamic instabilities, have been confused with full-load
levels and cause vibration trips. tests. Full-load tests do not necessarily duplicate field operating
Vacuum testing cannot determine whether aerodynamic in- pressures, and therefore do not ensure field operation will be
stabilities will be present since the impellers will not be
free of aero instabilities.

Best
Best Practice 3.26Practice 3.26
Surge protection system design e incorporate a dedicated It is found that this will:
bypass line intercooler in large systems to minimize the Reduce surge system response
volume of trapped gas for optimum surge system response. Require increased surge to surge control line margin
Using a dedicated surge line intercooler allows the surge line Reduce the compressor operating range
takeoff and check valve to be connected immediately after the dis- Possibly require an additional hot gas bypass line to prevent surge
charge flange. during emergency shutdowns
This design results in the smallest volume of gas that can re-enter
A recent gas booster project that used the aftercooler and not
the compressor during a surge cycle and provides the quickest surge
a dedicated surge line cooler required a large surge to surge control
system response.
line margin of 20% to prevent surge during emergency shutdowns
A fast system response allows the surge control line to be set close
(ESDs) thus reducing the operating range of the compressor.
to the actual surge line which results in the largest possible com-
pressor operating range. Benchmarks
Incorporate this design during the pre-FEED phase of the project to
This best practice was recently implemented for a gas booster
ensure implementation.
project and justified based on total cost (not having to use an ad-
Lessons Learned ditional hot gas bypass line) and reliability (minimum surge system
hardware).
Using the aftercooler for recycle gas heat removal, as
opposed to a dedicated surge line cooler, results in a large
volume of trapped gas.