Turbine Air Inlet Filter Pulse Cleaning


Thread Starter



My question is related to the pulse cleaning of turbine/generator inlet air filters. Our machine uses pulse cleaning with Dwyer control. In the dwyer dp gauge there are three pointers. One for actual dp across the inlet filters and 2 others for preset dp and high set dp.

I want to know how the preset dp and high set dp are used in the pulse cleaning process? Is the preset dp, the dp above which the pulse cleaning timer starts and stops when the actual dp reaches below it? Is the high set dp value used to over-ride the timer based cleaning? In manual cleaning mode, how does the reverse purging of filter continues?

What are the typical values of pre set and high set dp in dwyer gauge? Ours is a frame 6 gas turbine of 30 MW capacity.

Have you consulted the inlet filter systems instructions provided with the turbine-generator and auxiliaries?

Also, there is a document provided with nearly every GE-design heavy duty gas turbine named the 'Device Summary' and it usually contains all of the device settings (calibration settings) for the various devices provided with the turbine and auxiliaries (not usually including the generator protective relays or generator instrumentation (watt and VAr transducers; etc.)).

USUALLY--and every site and turbine is NOT the same--one of the gauge needles (the lower of the two settings) is adjusted to start pulsing the inlet filter solenoids to clean the filters when the differential pressure across the filters reaches a level determined by the filter manufacturer or inlet filter equipment supplier. This pulsing of the cleaning mechanism continues until the pressure drops below the first needle's setting. I have seen one of the pulse control printed circuit board manufacturers provide the ability to determine how many cycles the system will go through when started by a differential pressure switch.

The other needle's setting (the higher of the two settings) usually annunciates an alarm via the turbine control system of a high inlet filter differential pressure. And, should the inlet filter differential pressure increase even further, there is usually a separate differential pressure switch to initiate either an orderly, automatic shutdown of the turbine (an automatic STOP) or a turbine trip (which depends on the turbine packager's requirements).

There is often--but not always--a timer in the inlet filter control housing which can be used to start pulse cycle. Usually, that's done on a daily basis (once every day, for example), but some sites with very dusty environs or conditions use the timer for more frequent pulsing during each day. Some of the pulse control printed circuit boards can be set to complete one, two or more cycles of all of the cleaning solenoids when automatically activated by the timer.

I have never encountered a self-sequencing inlet filter cleaning system that started the timer when the filter differential pressure reached a certain value (the lower of the two needle settings, for example)--but that could be possible, and, I haven't seen every configuration of every auxiliary that's ever been produced.

But, the above are the typical basics of a self-cleaning turbine inlet air filter system.

The filter manufacturer usually defines the differential pressure at which the filters should be replaced. And, when supplied as part of a self-cleaning inlet filtration system the filter manufacturer will usually specify the differential pressure at which the pulsing of the cleaning system should begin. Since many turbine owner/operators buy filters from suppliers or manufacturers OTHER THAN the original filter manufacturer the filter manufacturer of the filters being used on any given turbine should be consulted for the proper setting of the device that initiates cleaning. As a suggestion, the setting for automatically initiating cleaning might be 50% or 67% of the differential pressure at which the filter manufacturer recommends replacement of the filters.

It should be noted that if turbine inlet air filters are not replaced when the differential pressure across them gets "high" (for example, above the upper needle's setting) it's possible for the suction created by the turbine's axial compressor can cause the inlet filter elements to rupture which can result in the ingestion of filter components and dirt and debris which had been trapped in the filter into the axial compressor, causing severe damage to the axial compressor. Also, on some sites there is/are implosion doors on the inlet ductwork which will open when the differential pressure gets very high and will allow unfiltered air and dirt and debris to enter the axial compressor. In the worst case, the inlet filter ductwork can implode--collapse inward.

Lastly, self-cleaning turbine inlet filter systems cannot remove all of the dirt and debris trapped on/by the filters. And, if the environment is very humid and the dust is very fine or of a particular nature (like cement dust, vapours from a nearby refinery or chemical plant, exhaust from the vehicles on a a nearby road/highway, etc.), the dust/dirt on the filters will adhere to the filters and cannot be removed by the pulsing of the filter cleaning system. At some point--and that point differs for every turbine installation based on ambient conditions and turbine operation (time; load)--the turbine inlet air filters must be replaced. Self-cleaning turbine inlet air filter cleaning systems are very effective in some cases, not very effective in some others. Generally, they increase the time between filter replacements, but they don't prevent replacing filters.
Thank you very much Mr. CSA.

I got the filter's operating parameters in the Device Summary of the manual.

The cleaning cycle starts at 650 Pa; stops at 450Pa; Alarm set (high) is 1500 Pa and Alarm setting (shut) is 2000 Pa.
A queston rises here:
<b>What initiates the "Air inlet filer dp high" alarm --- the high dp set point of Dwyer gauge or the 63TF-1 pressure switch?</b>

While Dwyer's pointer is adjustable, 63TF-1 has a fixed value.

Hmmm.... Without being able to see the electrical schematic for the turbine inlet air filter self-cleaning control enclosure, and based on the information provided I would say the differential pressure gauge may be used for starting the pulsing (at 600 Pa) and stopping the pulsing (at 450 Pa), and differential pressure switches will be used for alarm and shutdown.

But, that's just a SWAG (Scientific Wild-Arsed Guess) based on the information provided. Usually the devices are all labeled in the turbine inlet air filter control enclosure (with stickers on the backpanel, or stuck to the devices themselves).

But, there should be an electrical schematic provided by the inlet air filter self-cleaning manufacturer which would be helpful in determining which device does what function.

Again, not every GE-design Frame 6B heavy duty gas turbine and auxiliaries are exactly alike. They're usually similar, but not exactly alike.

Hope this helps!
Thank you Mr. CSA,
Going through the manual, it is verified that the Dwyer's gage set pointers are being used for starting and stopping the pulse cleaning system. 63TF-1 is used for alarm annunciation and 63TF-2A/2B for shut down.

About manual operation of cleaning system, the manual says that the cleaning starts immediately irrespective of the dwyer's set point.
I think that in manual, the cleaning progress is timer based and only the start by dwyer's gage is bypassed.

The operation of dust blowing from the hopper is still not clear. Perhaps dust is blown out of the hopper at the end of the cleaning cycle when the actual dp falls below the stop set-point.

Rest, i will confirm the whole conclusions in field and update here soon.

This was my first interaction in control.com though i was following it for 2-3 years now. It was very helpful indeed.

Thanks again.

Hmmm.... Glad to see you've RTFM (Read The Fine Manual). It's unfortunate that most technical manuals for GE-design heavy duty gas turbines and auxiliaries are written by technicians/engineers, and sometimes it's necessary to RTFM more than once to get a good sense of the equipment, its operation and troubleshooting. It's also quite common to find a procedure in a manual with steps 1, 2, 3, 4, 2A, 2B, 5, 6, 3A, etc. In other words, do this, then that, then that, then that, oh--but before Step 3 make sure to do this and this, then back to next step, .... So, it's sometimes necessary to read the entire procedure at least once, and then write one's own procedure. Actually, it's ALWAYS best to RTFM at least once, and if developing a procedure, or intending to follow a procedure in the manual, it's best to write one's own procedure so that one is confident in what has to be done and the order of the steps.

The manual self-cleaning push-button (sometimes labeled START)--as opposed to the instruction manual--is used to bypass everything: the timer, the gauge, any other pulse start method. Pushing it means a human wants to initiate the pulsing. And, depending on the manufacturer of the self-cleaning control printed circuit board(s), the cleaning will go through at least one cycle, or two or three if there are settings/configuration for that. Some self-cleaning control systems had both the "start" methods paralleled at one input--the MANUAL START, and the gauge start.

Ohhh, the self-cleaning inlet filter system has a hopper. Does it have a screw-type auger in the hopper? My experience with these systems which include a hopper/auger is that they don't work well except in certain circumstances. The ambient has to be very dry (low humidity); the dust has to be a little "coarse" (i.e, not too fine); there must be little or no ambient wind when pulsing the filters to clean them. And, the hopper needs to be emptied when there is no wind and into some kind of closed receptacle. (Oh, did the turbine package supplier forget the closed receptacle and "hose" to connect it to the outlet of the hopper? Then nothing has changed in decades.)

Thanks for the feedback! Best of luck with commissioning/troubleshooting the system.
Dear CSA,

We have GE Frame 5 Gas Turbines.

The pulse air cleaning system is similar to what you described above. We have 3 pressure switches 63TF-1 (Set. 11.3mbar) & 63TF-2A/2B (Set. 30 and 32 mbar). The first one initiates an alarm, while the turbine trips on 2 out of 3. We are located in a remote desert where humidity can rise during winter. The turbine tripped few times when there was a fog and humidity rose, since the differential pressure exceeds 32mbar.

What will happen if we increase the setting of 63TF-2A/2B?

What is main purpose of that protection?

Are 63TF-2A/2B installed on all gas turbines?

Does it have anything to do with air flow rate and surge?

From my first response to this post, if the inlet filter differential is allowed to get too high bad things can happen. If the unit has implosion doors the implosion doors will open--which will allow unfiltered air to enter the axial compressor. (Limit switches should either trip the unit, or initiate an automatic shutdown when the implosion doors open--to prevent continuous ingestion of unfiltered air into the axial compressor.)

If there are no implosion doors (also, usually a purchased option, by the way), and the inlet filter differential gets too high the metal walls of the inlet duct (downstream of the filter house) can implode. Vacuum is a very powerful thing, and most of the inlet duct work is not reinforced to prevent buckling due to implosion. I have never experienced an imploded inlet duct, but I have seen pictures, and it looks like someone dropped a very heavy weight on the duct work, crushing it--and allowing unfiltered air to enter the compressor.

The function of the implosion doors is to try to prevent a high inlet filter differential from causing the inlet duct work to implode. The operation of these doors is very important, and often overlooked during maintenance outages.... Without the implosion doors (we don't know if your inlet structure is equipped with them or not), there is a danger of implosion.

Another bad thing which can happen when the inlet filter differential gets too high is that the filter elements can rupture and allow all of the dirt which was trapped in the filter to enter the axial compressor, as well as unfiltered air through the ruptured filter. This is also bad, though not as bad as imploding the inlet filter duct, or opening the implosion doors (if so equipped).

The solution to your problem is to work with filter suppliers to find a different filter material than the one you are currently using that doesn't swell when it gets damp from the humidity. There are many different filter materials which can be used--and the new material may be better in many respects to the old one.

I'm surprised to hear that 63TF-2A & -B are not set to the same value; that seems a little odd.

I would be hesitant to increase the setting of either--again because of the danger of rupturing filters, opening the implosion doors (if so equipped), or imploding the inlet duct. The best solution is to find a different filter material that doesn't well when it gets damp.

Hope this helps!
The three pressure switches provide implosion protection. If the differential pressure drop in the transition is greater than the calculated value stated on the top-level drawing, then an alarm (63TF-1) would be noted by the DCS. If the differential pressure drop continues to increase the a shutdown (63TF-2A -2B) then a shutdown would be noted by the DCS. The set points should not deviate from the specified values on the filter house drawing.
Dear CSA,

I think turbine shutdown limit on high filter DP at 30 mbar (12 "WC) is too high. It should be around 15 mbar (6 'WC).
engr 14,

It does seem high, but all inlet filters are NOT created equal(ly). Some can withstand a higher differential, while others can not. It all depends on the type of filter, the type of airborne contaminants (including humidity, and hydrocarbon vapours, if any), and the construction of the inlet filter duct work (because it also has to be able to withstand a high differential (vacuum).

High differential pressures across the filter will also, typically, cause the performance of the unit to be less than normal for any given condition. (It's like putting a surgical or dust mask on one's face, and then trying to run fast for some distance, breathing through the mask. It doesn't work very well.)

But, just to say some value is "high" (or "low") does not take into account all the variables which can be a factor in deciding what the alarm- and trip/shutdown values should be.

Hope this helps!!