Compressor Bleed Valve Fail Safe position


Thread Starter


I have worked in two types of GT 501 frame gas turbines and have noticed that the fail safe position of their compressor bleed valves are different; one is fail open and the other one is fail close. Can somebody share their thoughts on what is the better comp bleed valves fail safe position with respect to GT protection?

Many thanks in advance.

Better position is fails open. what is normal?

when we start or stop machine then there are more chances of surging.
if during load operation Bleed valves open then only affect is MW and efficiency reduced. but if that is fail close, then during shutdown or startup can cause trouble to machine
Hello JCP,

I have to say one thing first: I never worked on Frame-5 engines. So I can only give you some general information, although it woud be a very big surprise for me to hear that the fail-safe position of a bleed valve connected to some stage of an axial compressor for gas turbine is not OPEN.

Basically the purpose of bleed-valves is protection against rotating-stall and/or surge of the axial compressor, especially concerning the last stages. Whenever the gas-turbine is not running on nominal design speed (which actually is maximum speed) or is not operating under ISO-Conditions (regarding ambient temperature and ambient pressure and above sea level) the gas turbine is in off-design mode. As ambient pressure for a specific site-location (is more or less) fixed, the ambient temperature of course changes throughout the year and has an impact on density of compressed air and therefore influences the power-demand, the discharge pressure and the discharge temperature of the compressor. But even if you meet the ISO-conditions of the axial compressor having exactly design ambient temperature and ambient design pressure as a very big coincidence, which should only happen for a few minutes over the year, there is still a problem with speed. Because speed changes all the time, maybe with the exception of single-shaft turbines running in steady-state conditions for generator application. Two shaft engines for mechanical-drives have a range of speed for both, the gas-generator (axial-compressor) and power-turbine, as the load (for example Natural Gas Compressors) needs variable speed to control the flow and discharge pressure of natural gas. But any speed lower than maximum (design) speed means off-design operation of the full gas-turbine-compressor set. Beside the permitted range of speed in normal operation of course any gas-turbine, regardless of if it has one, two or even three shafts, has to be "safely guided" through its safe band of the compressor's map from zero-speed to operational speed when starting the engine or vice-versa when stopping the engine. And on its way from zero-speed to operational speed there are a lot of things happening in an axial compressor as it is quite a complex assembly of mechanical components from an aerodynamic and thermodynamic point of view. Experts speak of things like "critical speeds", "local blade oscillations", "stall-line", "surge-line" and even more. I definitely can not explain to you all the details of critical flow-conditions of an axial compressor as I am not a mechanical engineer, but "only" a control engineer. But what I can tell you is that the surge-line is a line of minimum flow and the stall-line is a line of maximum flow, both depending on differential-pressure across the axial-compressor as well. In addition there is a steady-state point of view and a transitional/dynamic point of view. From a steady-state (static) point of view the last stages of most axial-compressors are not able to transport the incoming compressed air from previous stages until the compressor reaches a minimum speed. So what design engineers often decide to do is to use "variable geometries" to reduce the air reaching the last stages of the compressor. A variable geometry can be either variable inlet guide vanes (IGVs) or (now coming back to your question) bleed-valves, being connected to the stage in front of the critical stages allowing the excess of air to bypass to atmosphere to prevent stall of the last stages. When the minimum speed of the engine is reached the last stages are ready to compress all the air which is coming in and the bleed valves are closed. In addition, beside IGV and bleed-valve control, it's all about fuel-control when accelerating/starting or decelerating/stopping the engine. What you have to know is the dynamic line on which the operating point of an axial compressor moves through the compressor's flow map. While accelerating the operating point tends to move more horizontally to the right first (in direction of higher flow) and after some time also the differential pressure comes up moving the operating point upwards.

While decelerating the operation point also first moves horizontally to the left (in direction of lower flow) and after some time also the pressure drops moving the operating point downwards.

So in the first case the operating point is converging the left operation limit, which is the stall line. In the second case the operating point tends to cross the right operation limit, which is the surge line.

So the target of the ACCEL and DECEL fuel control schedules is to limit the change of speed per second to provide safe margins to both operation limits, moving the operating point through this safe gap between stall- and surge-line starting after light-off is detected from purge-speed to destination speed (Idle-speed or any speed setpoint above until maximum or design speed).

On some engines you will only find inlet guide vanes controlling the overall permitted amount of air (defined by the maximum air which can be compressed by the last stages) by adjusting the angle of airflow at compressors inlet (bellmouth). On some engines you can even find both IGVs *AND* Bleed-valves. And on some engines you can find position-controllable bleed valves which are not only for protection of the axial-compressor but also for flame-temperature control on DLE engines due to emission requirements. IGVs are often used on higher frame engine classes for exhaust-temperature control due to efficiency improvements.

So in most cases there are a lot of different functions and protection circuits involved and combined in override-control-loops using the same actuators. But the safe or fail-safe position of a bleed-valve on an axial-compressor (as ar as I know from GE's LM-engines, GE's PGT10, SOLAR's CENTAUR, MARS, TITAN and Hispano-Suiza's/MAN-Turbo's THM-engines) definitly should be the open-position (full bleed of air to atmosphere or to exhaust stack) in the meaning of highest protection against rotating-stall.

The safe or fail-safe postion of IGV's should be full-closed for highest protection of the compressor against surge and rotating-stall.

So never close a bleed-valve, before minimum speed of the compressor (the last stages) is reached!

And never open the IGVs more than defined by the control schedule!

This is nearly all I can tell you about bleed-vales and their safe-positions. When you say that you have found two bleed-valves on your engine having different safe positions then I would strongly advise you to have a close look at the P&IDs and on the connectios of valves&pipes at the engine again. Because sometimes things are not as they seem to be! For example GE's PGT10 engine has only ONE bleed-valve but two valves which in the documentation refer to be "bleed-valves". BUT: One of these two so called bleed-valves (20CB) is just a solenoid which switches control-air to another valve (the "real" bleed-valve), moving it's piston in such way that the "real" bleed-valve is closing. In such constellation the safe position of this "real" bleed valve is OPEN and the safe, de-energized position of the solenoid (20CB) is CLOSE to safely cut-off supply of control-air to the "real" bleed-valve, causing it to OPEN immediately protecting the axial compressor against stall under all circumstances.

Hope this helps!

I believe the originator was referring to Circle-Bar-W (Westinghouse) 501 heavy-duty gas turbines.

Perhaps JCP could clarify that for us all.

To all who replied to this post thank you very much for sharing your thoughts; they are so informative. To clarify some information, I am referring to a Mitsubishi M501G engine which has a fail open bleed valves and a Siemens Westinghouse W501D5A engine which has fail close bleed valves configuration.

I also believe that open should be the safe position of comp bleed valves; but I am also interested to know if someone out there knows any good reason in having a fail close bleed valves over a fail open ones.

Have you looked at the Instruction Manuals and the P&IDs provided with the two units? And have you consulted the parts list and drawings to see what the two manufacturers originally provided with the two units?

Because, I would have to believe that someone would be seriously endangering a machine and a serious possible loss of production by replacing normally open compressor bleed valves (I presume there is more than one bleed valve per unit) with normally closed compressor bleed valves, or vice versa. That's a pretty gutsy move to make without some kind of peer review or without consulting the manufacturer.

Perhaps the manufacturer issued some kind of safety- or informational bulletin to replace the valves?

But one would think someone at the plant would know.

Please let us know what you find out.
Dear JCP,

One question please:
I haven't worked on both engines you mentioned. But, are these two types of gas-turbine in some way comparable at all?Is it the same type of engine only manufactured by two different suppliers/packagers (MITSUBISHI/SIEMENS)? Or is one of them a DLE-engine and the other one a non-DLD-type?Or is the only common thing about these two units that they are, well, gas-turbines?

Please also reply to CSA's question and give us the whole history and the big-picture first. What happend? Who did what and why?Are these two valves you are refering to exactly installed at the same point when you have a look at the P&IDs?

Please give us more background. I'm sure we can help you if only you support us with enough details.