Has anyone tried a P.M approach to the IGV servo sticking problem by going in the calibration mode and stroking the valve once a week? We have two 7FA that we run every couple of weeks during the spring and early fall months that are prone to sticking during these down times. We are trying to come up with a short term approach for this problem.

 

This problem is related to two issues: temperature and lube oil formulation.

I'm presuming the servo you're having the most trouble with is the IGV servo.... That's just a sway, but it seems to be the most common complaint. The gas valve servos also seem to be problematic at some sites, but that's also when they indiscriminately run the gas fuel module vent fans in colder ambients....

I have yet to hear of anyone insulating and heat-tracing the hydraulic/trip oil lines to the various valves. I would think that would go a long way towards helping the temperature portion of the problem.

It's also pretty clear that the compartment space heaters used in many F-class turbine compartments are woefully inadequate and poorly placed. Sure, they prevent freezing of cooling water- and water injection lines, but they don't do a good job of helping keep the L.O./hydraulic/trip oil lines warm enough to prevent viscosity issues.

When all of the servo-operated devices were basically on the turbine- and/or accessory base (which was also before turbine lube oil formulations changed), these kinds of problems weren't so prevalent. With the change to off-base skids and compartment and device locations, this problem has grown.

The lube oil formulation has been covered in previous threads on control.com. The oil companies have collectively improved the lubricating properties of 30 weight turbine oil, but for those turbines that use the same fluid for hydraulic systems it seems to be causing problems with servo-valves, and trip/dump valves and -solenoids in some cases.

But, combined, the temperature and formulation issue seem to be the root causes of the problem. GE have some technical information about the L.O. formulation problem. Their primary mitigation solution is to do some very expensive "polishing" (I forget exactly what it's called). Castrol has developed an oil for turbine applications that use L.O. as hydraulic fluid which has proven to be extremely helpful in reducing servo-valve problems. It's widely used in the UK and Europe.

Your suggestion of an automated or scheduled "stroking" of servo-operated devices does nothing to address the root cause of the problem. It's simply a band-aid, and a poor one at that.

But if it works at your site, go for it! Every site is a little bit different. Turbines all suck and squeeze and burn and blow; but the auxiliaries and layout are a little different at every site.

 

The problem with the varnishing is well documented with the 7Fa fleet. I'm trying see if anyone has had any success with this method. Part of the problem is there is little or no flow through the servo when there is no demand for movement. You can polish the oil in reservoir all you want but you will have little effect on the IGV servo circuit, without putting the flush blocks on for some period of time. We have considered putting heat tracing on this circuit to see what results we get. The lack of flow still comes into play is the reason we wanted to stroke IGV to see if we get any positive results.

 

I haven't done a lot of research on the mechanism of varnishing, but why does low temperature make varnishing worse?

There have been a number of peaking or load following F-class sites that report stroking of the IGV actuator prior to scheduled operation. It usually involves forcing logic and using AutoCalibrate, both of which most operators are not pleased about (it's a change from their normal routine, and power plant operators are loathe to change their routine, which is another operational issue to consider).

And, why does exercising the servo in low ambients result in better operation? If it were strictly related to varnishing, then exercising the servo/actuator wouldn't seem to be the answer to the problem.

And to the well-documented problem of varnishing, I'm curious to know why people aren't using the Castrol blend, or asking their L.O. supplier to make a similar blend? With the proliferation of F-class machines, it would seem there would be a pretty big market for this oil formulation.

Again, until the F-class turbines came on the scene and servos were "remotely" located from the L.O. tank this problem was not nearly as severe as it has become--particularly with the IGV servo. And, I believe that's directly attributable to the length of the hydraulic/Trip Oil feed/drain piping, the low flows through the piping as you have mentioned, and the temperatures in the Turbine Compartment.

Where is(are) the Turbine Compartment space heater(s) located in the Turbine Compartment at your site? In the upper area of the Turbine Compartment, or in the lower area of the Turbine Compartment? Is the air flow in the Turbine Compartment uniform when the space heaters are operating--meaning, is every area of the compartment getting a reasonably uniform warming air flow?

Have you put some temperature sensors on the underside of the hydraulic/Trip Oil piping in the Turbine Compartment to monitor the temperature during the colder ambients? (Isn't the IGV hydraulic/Trip Oil feed/drain piping in the lower portions of the turbine compartment, where the temperature would be lowest?)

Again, there's a root problem that needs to be solved. It involves low ambients and oil formulation--and as you have indicated, low flows, also. If it were strictly a varnishing problem, then why wouldn't it be a problem at any ambient?

Another thing to remember is that the servos in most F-class applications are "exposed", meaning they are visible to the naked eye--and to ambient conditions. In non-F-class turbine designs, the servos were almost always "in" the top of the L.O tank (with the exception of the IGV servo, which was in a much more compact turbine compartment. I think insulating/heat tracing the servo might be part of the solution, particularly the IGV servo. To this end, try putting a temperature sensor on the servo itself and monitoring this temperature over time and operating conditions.

If you do insulate and heat trace the piping and servo, please write back to let us know what the results are.

 

Here is an interesting Internet search engine result (remove any spaces inserted by the website's software):

http://www.turbomachinerymag.com/white papers/Seaworthy_whitepaper.pdf

It doesn't explain why "varnishing" is worse at low temperature. It does indicate the formation of varnish is a function of extreme (high) temperatures. It's possible that low temperatures and low flows can result in the collection of varnish "residues".

I'm sure there's lots more varnish info available on the World Wide Web.

I'm just suggesting that it's best to try to understand and solve the root problem, once and for all, then to institute work-arounds which inconvenience people (operators) and operations.

Again, please write back to let us know how you progress with this problem.

 

Speaking of having operators forcing logic, I was contacted just this morning to try to help understand why the IGVs wouldn't open during starting on an F-class machine with a Mark V turbine control system.

Turns out the site has a weekly policy of having the operators manually stroke the IGVs at 4:00 am every Friday. Allegedly, since instituting this policy it's claimed the IGVs have never failed to operate properly, so everyone at the site was certain there was nothing wrong with the IGV servo.

Manually stroking the IGVs involves forcing logic signals L20TV1X, and this morning instead of unforcing L20TV1X they forced it to zero and left it forced at zero. The next shift came in, paid no attention to the alarm "Logic Forcing Detected", and proceeded to START the unit, and the IGVs wouldn't move.

The on-site I&C tech called in sick, so the site was frantically trying to understand what had happened.

It took about an hour to get the details by phone and find the forced logic signal.

The unit successfully started, but by that time, the peak had past and the plant missed their window for the morning.

Needless to say, the Plant Manager was pretty upset--at everyone and anyone. Including GE and everyone who was "involved" (yours truly included, and I've never been on site!).

It also turns out the site has no written procedure or check-list for this activity.

This is precisely what can happen when work-arounds are instituted instead of solving the root problem(s).

Varnishing occurring during colder ambients. I question this is the culprit. I believe the fact that the new formulations being used for L.O. cause the oil to varnish faster than before is more likely to blame. Maybe cold ambients cause varnish to "collect" in servos faster, but stroking valves isn't going to <b>solve</b> that.

 

I have on question. Was the "flush oil" used during the installation of the turbine re-used?

Sometime vendors convince others that the "flush oil" can be filtered through an "special filter" and used as turbine oil.

 

Interesting. Can I find a photo or drawing of the IGV you are talking about? Our company has developed an electronic governor based onto a SCADAPack 32 and in 2009 replaced an old English Electric mechanical hydraulic unit we inherited when we bought the plant, which was hunting badly, and we had no knowledge. We also installed a new pilot servo with new Rexroth proportional valve. The new valve is sluggish. The oil to the old distributing valve is warm, but the new pipework to new proportional valve and pilot servo is dead cold. We have a free standing heater. The engineer has suggested heat trace. I suggested a bleed valve placed between two P and T port blanks on the manifold below the proportional valve. But oil in the pilot servo will still be dead cold. We increased dither settings, but I don't like this for a wear perspective. The space heater remains until present

 

You may be interested in either the CRV plate (Oil circulation system)supplied by Thomassen or one of the Balanced Charge Agglomeration filter systems (Isopur,Kleentek). Both seem to have good reputations in reducing servo sticking due to oil lacquering. If you want to drop me an Email at [email protected], I can send you some info. on both.

 

Regarding the temperature effect previously mentioned, see the temperature vs varnish solubility flick at

One can't fool mother nature.

Regarding root causes: using the same oil+sump+system for both bearing lubrication and control oil is the root cause of the noted reliability issue. This shared-design is highly compromised due to inherent conflicts between the two (lubrication vs controls) independent sets of design requirements. The only reason for sharing the systems is cost reduction for the turbine maker. Turbine makers that design independent lube oil and control oil systems avoid failing servo control valve performance due to varnish because control oil system maintenance is documented, easy and affordable (small sumps).

 

Until the oil companies changed their formulations, this was not a problem.

Until servo-operated devices were located "remotely" from the L.O. tank via uninsulated lines in poorly heated compartments, this was not a problem.

There are many logical reasons for using the same fluid for lubrication as well as hydraulic applications, several of which PB mentioned. Another is simplification of piping, sumps, reduction of the number of filter spares, and a reduction in the number and types of oils which must be kept on hand as spares. Reducing piping and sumps is not simply a bonus for the manufacturer, it also has many pluses for the owner/operator.

This is just a bump in the evolutionary road which must be dealt with. Stroking hydraulic actuators, manually or automatically, is <b>NOT</b> the solution to the problem(s). And, it has many drawbacks.

I was also reminded of another site which reached FSNL and could not automatically synchronize the generator breaker. The operators had forced the logic signals for the Aux. L.O. and Aux. Hyd. Pumps ON during the manual stroking "procedure" instituted for this very same issue and then left them forced ON, instead of unforcing them. The application code in the Speedtronic blocks automatic synchronization if the Aux. Pumps are running (this particular unit had Accessory Gear-driven L.O. and Hydraulic pumps.)

And, again, this failure to synchronize was attributed (incorrectly) to a logic problem, not an operator failure to follow procedure (in that case, there was actually a check-list procedure, and the operators had checked that the signals were unforced--but, in fact, they weren't).

Root cause analysis and solution is the answer to reliable operation every time.