In this case study we have a turbine exhaust surface condenser with a steam ejector air removal venting system that is having performance issues. The air removal system was a two stage twin element design, meaning one first stage ejector and one second stage ejector per element or train. Therefore, only one 100% train is intended to operate at a time and one train is a spare. The turbine exhaust condenser was seeing an operating vacuum pressure poorer than the design of 3“HgA, operating at a pressure of 6.2“HgA. A pressure and temperature survey was performed on the entire system multiple times under a yearly service agreement. The following was observed from the survey data.
The air removal system was operating with both 100% ejector trains in service in order to maintain a reasonable inlet operating pressure on the turbine exhaust condenser. In addition to the main air removal system, the hogging ejector was in operation. The hogger ejector is meant to only operate at start up, pulling the system from atmospheric pressure to 10”HgA within 30 minutes per HEI (Heat Exchange Institute). The hogger is not intended to operate all the time. Operating the hogger ejector all the time will result in additional operating cost, as it uses significantly more steam then the first and second stage ejectors. In some cases, the hogger will also limit the vacuum level the condenser operates at. Refer to a previous article
Proper Use of Hogging Ejectors for more details on the proper used of hogger ejectors and estimated cost to run a hogger yearly.
Before looking at the Air Leakage Meter, there are two important details to be aware of:
- Make sure the packing foam has been removed from inside the meter. Many times we find the foam has been left in place. This foam is placed in the meter to protect it during shipment.
- The ALM will give an accurate air reading only when the hogger jet is off and isolated. When the hogger is operating, most of the air will go through the hogger, so the ALM will indicate a much lower (and false) reading.
The design air leakage for the system was 45 lbs/hr. During the first visit, the air leakage meter was maxing out the 0-120 lbs/hr air leakage meter.
With the air leakage being so high, a helium leak test was performed and the following leaks were found:
- A relief valve was found to be leaking
- The turbine steam outlet to condenser piping flange was found to be leaking
- The pipe union nearest the condenser on the loop seal was leaking
- A pipe union on the level transmitter was leaking.
Once these items were sealed additional pressure and temperature surveys were completed at later dates. During the next visit the air leakage meter was reading 100 #/hr, however the hogger ejector was in operation. Therefore, the reading of the air leakage meter at the aftercondenser would not be fully accurate as the air leakage is split between the hogger and the air removal system. When the hogger was removed from service, the air leakage meter went to the max 120 lbs/hr limit and the turbine exhaust inlet pressure degraded. The turbine exhaust surface condenser was operating over three times worse than design at lower loads at 15.5"HgA. The system was found to have turbine gland seal leaks, causing the air leakage into the system that was not identified in the first helium leak test. In this case multiple turbines were entering the condenser and one was not in operation at this survey causing the turbine glad seal leak.
Another observation was the inlet temperature to the ejectors and hogger were hotter than the outlet temperature of the main turbine exhaust surface condenser. This indicates that steam was being pulled backwards through one or more of the ejector suction connections. At this time one of the first stage ejectors was removed from service and the turbine exhaust surface condenser improved to 185 mmHgA, but still not at the design of 3”HgA.
The next visit all the turbines were operating and the turbine exhaust condenser was operating at 2.84”HgA at about 80% of the design steam load. This seems better except the air leakage rate was reading 45 lbs/hr with the hogger ejector operating using additional motive steam. When the hogging ejector was removed from service, leaving only the two second stage ejectors in operation the air leakage rate increased to 85 lbs/hr, when the design is 45 lbs/hr. Also, the turbine exhaust pressure degraded to 3.03”HgA. At this time one first stage ejector was placed in operation and the turbine exhaust inlet pressure improved to 2.84”HgA. The other first stage ejector was placed in operate to have both trains operating, but the vacuum degraded to 3.27”HgA and it was quickly removed from service. The system was left with one first stage ejector in operation and two second stage ejectors in operation due to the higher than design air leakage rate. It was later found the remaining air leaks were in the turbine glad seals and need to be addressed to additionally improve the performance.
This case study shows that air leaks are difficult to find and it may require an ongoing effort to find and eliminate all the air leaks. This is especially the case for new installations or after turnarounds where equipment has been altered. It’s more likely after a new installation or work were equipment has been disturbed to have flanges, valves, glands, seals be loose and not sealed fully. A good way to identify there is an air leak is to look at the last stage ejector. The last stage ejector will become overloaded with the non-condensable gases when a large air leak is present. In this case study both second stage ejectors are still in operations to overcome the higher air load as the air leakage is still higher than design.
