I would like to get some idea for our trouble-shooting by discussing the stress analysis method of rotating equipment piping with members.

We have trouble in a turbine driven compressor piping. The coupling bolts between compressor and turbine shaft were broken as soon as start-up of turbine in few seconds. Of course, vendor and our mechanical team have checked all possible sources one by one. After long deliberation, vendor reported the most possible source of this failure is the excessive nozzle loads from external piping and they pointed out the friction loads. In my experience, I do not think the piping is a source of this trouble. Although piping loads are higher than allowable limit, these excessive loads will only cause higher vibration level than allowable limit, other than coupling failure.
We thoroughly checked the stress analysis again and the installation conditions such as piping supports, dimension, and weights at site. The deviations between original design basis and installation are minor in view of the stress analysis. The nozzle loads are well within NEMA allowable limit for normal operation condition by neglecting friction. We did not consider a frictional effect for normal operating condition and we used hot-setting -spring other than cold-setting for the spring supports adjacent nozzles. I would not like to discuss the setting method of spring since this may be controversial topic. Fortunately the differences of nozzle load between two setting methods are minor and both are within allowable limit.

The outstanding problem is friction loads. Turbine inlet and outlet piping(inlet 10”, 480 Celsius, outlet 14” 200 C) have spring supports(F-type) for first three points with Graphite sliding assembly on the top of spring load flange respectively and the rest are resting(shoe-type) support on beam. The result of Caesar II analysis by including friction factor(0.3) only for shoe(resting) support shows to exceed 200% allowable loads based on NEMA.
Our practices for the stress analysis of rotating equipment piping are as below;

1) Friction is considered as a short term transient load. Allowable load limits on equipment may be increased by a factor of 1.5 when considering normal loading plus friction loads.
In my opinion, generally the piping connected to turbine nozzle is fully hot with proper thermal expansion during compressor turbine warming-up. During this warming-up stage, the friction loads will be imposed to nozzle but not results in turbine failure, even though friction loads higher than 150%. Once thermal expansion occurs during warm-up, the friction effect is negligible during normal operation.

2) The rotating equipment allowable loads are the loads permissible for long term continuous operation. These loads are therefore compared with the piping loads calculated based on normal operating temperature. However unusual design or flexibility temperature conditions, such as steam-out, loss of fluid, seismic event, and so forth which occur mostly when the rotating equipment are not operating, estimated by simple temperature ratio, exceeds the allowable loads by more than 50%.

3) The nozzle loads may be calculated using the modulus of elasticity for the operating condition.

Based on above,
a) Are above our practices reasonable ?
b) Could the friction included loads really cause this failure ?
c) Is piping stress engineer responsible for ensuring the safety of turbine during warm-up condition ?

Our final intent is to show and ensure the external loads from piping did not result in this failure. Also we want to get out of trouble.

I hope that many members advise good experiences to us for this topic.


[ June 14, 2001: Message edited by: Soon Ryang, WEE ]