dual fuel diesel/gas turbine maintenance

[vmac100]

hi good people,compliments of the season and good work you have been doing here.
please i need some assistance with my project.i am trying to design a maintenance programme for our diesel/gas turbine generators
the turbines were installed and commissioned in 1981.they are Ruston TB5400,3MW generators.they were initially just diesel fuelled,but wrer upgraded to use gas in 1992.
when i joined the company,i observed that the generators were producing 2.3MW,average and 2.6MW,peak.this to me is unacceptable.and i took it as my first project to get a maintenance program that will increase the reliability and maintenability and hence availability of the turbine generators.also,whenever there is gas supply failure and we run the turbines on diesel,if gas supply is restored,there is always the need to change the burners and clean them of carbon accumulation before we can start them on gas.this takes close to five hours,in which time our consumersare left without power supply unless we use the stand-by generators,which is expensive,considering the cost of diesel.
what i need is information from studies and previous experience,and possibly applied maintenance programs.
thanks a whole lot

 

[vmac100]

also,i will like to know if proper maintenance can help improve the power supply of these generators from the present 2.4MW(which they say is the maximumdue to the age of the turbines)

 

[Romefu12]

For starters there are many variables to achieving greater efficiencies and reliability from your current engines. Greater output of turbo-generators can be achieved in a number of ways depending on initial design, operating environment, and fuel quality including;

1. Increasing airflow through the compressor by using more efficient air filtering or an upgraded compressor blade design to increase airflow.
2. Increasing air density through a colder air charge.
3. Water injection to increase air density.
4. Mixing of DI water with diesel fuel to create an emulsion for better combustion.
5. Choosing a diesel fuel of better quality. (Less sulfur content)
6. Reviewing the loading, unloading, starting and shutdown procedures for possible modification to operating parameters to alleviate some of the burner fouling that is occurring.
7. Request the assistance of the manufacturer or 3rd party suppliers on possible upgraded components including burner assemblies, compressor blades etc.


Not knowing all of the details necessary for a proper answer including operating environment, fuel quality, firing temperatures, compressor design flow as well as not being familiar with the initial design of these particular engines I can only lend some ideas to help your situation. In my opinion I would strongly recommend reading the manufacturers literature and look for ways of improving the operating parameters. With a few adjustments here and there you may be able to squeeze a few more kilowatts to improve your numbers. Regular inspections and scheduled maintenance would also greatly improve your reliability and efficiency.

Hope some of this helps.

Romefu12

 

[vmac100]

thanks Romefu12,that sure would help.however,i would like you to help me with some suggestions as to the operating parameters that you think should be paid particular attention to.
as for the maintenance and inspection schedule,i have already devised a time-based maintenance plan.what i require is additional input and advice.you know one can never have too much of information.
thanks,and i hope to get more from you

 

[Romefu12]

Not knowing all of the instrumentation, annunciation, and control parameters that are actually monitored in your engines (as they came with different levels of technology back then) I would need to know what upgrades, if any have been performed and what parameters are actually monitored. The older PLC’s have limited capabilities when compared to today’s data acquisition and performance monitoring capabilities.

Some areas of my concern would be range of operational ambient air temperatures and humidity, fuel pressures and temperatures (Diesel and Gas), fuel regulator flow monitoring and response times, Fuel valve operation and response times, power turbine exhaust temperatures in as many zones as possible, component corrosion, insulation condition. The rest would depend on what the current capabilities of your controls could handle. 3rd party companies offer TBC coated blades for increased life cycles, corrosion resistant flame tubes and burners for cleanliness and longer life cycles, improved flow filtering medium and compressor washing equipment for improved compressor air flow and fuel additives to control fuel quality. During future outages I would consider upgrading components to extend the life cycles.

Finally, your maintenance schedule should not be based on a time-only frequency but rather by a concept of an Equivalent Operating Hour frequency. (EOH) Most all Gas Turbine / Jet Engine manufacturers use this equation of EOH to evaluate maintenance and component life cycle assessment and replacement. This equation, which can be quite complicated, varies slightly by manufacturer and incorporates events such as the number of starts and stops, actual operating hours loaded and unloaded, unit trips, load rejection, failed starts, and a host of other parameters depending on the manufacturer.

To evaluate your particular situation I would need the whole picture in front of me and some time to study all the data available. Take your time and don’t assume anything. I have seen simple things like fuel valves and regulators cause many problems with operation. Sounds like an interesting project although in the end it depends on the amount of time and money you want to spend and what the future, if any, hold for these particular engines. I would think it would be wise to complete a cost benefit analysis before making any major investments. After all I believe you are speaking of only making up .6 MW.

Good Luck

Romefu12

 

[vmac100]

Thanks Romefu12,
actually,0.6MW sure means alot.if i can get 0.4 with three of our four turbines(using the N+1 philosophy),that will be 1.2MW! and there is an office area that we supply that has a peak load of 2.2MW during the weekdays and 1.6MW at weekends.presently,we are using the local utility to supply this area and this is costing us 8.50/kWhr.using or own generation,it costs us 5.69/kWhr.now if we can harness this extra 1.2MW and use the turbines and the stand-alone diesel generator sets,which are a part of our own generation,we will be saving 2.81/kWhr!i guess you see now that cost-wise it is worth the effort.
also,we plan on installing two 25MW dual fuel diesel/gas turbines by 2007.then we will have no need for the present turbines,right?what i think and i have told my boss is that there sure is supposed to be a salvage value for each of these units.i have confirmed same from Siemens.what i do know is that,even at the age of these units,if a customer comes and sees a functional,reliable,maintenable and available unit,he sure would pay more.
so you can see that what ever expenses we incur will eventually be repaid.
as for the time-based maintenance program i have,i have actually thought about that equivalent operating hour frequency.however,it was just a concept to me and i didn't really give it such a definitive name.what i was thinking as a part of the study,is to collect data on all the critical and some of the vital components,study their frequency of failure, and posit as a part of the maintenance plan the routine of replacing them before their 'failure time'.
as for the controls,i am still looking to study them.they sure are obsolete.normally the units are supposed to automatically start on one other fuel when there is failure or restoration of the other.however,that is not the case.
i will get you as much of the information you need to have a wider picture.surely,it is a very interesting project

 

[vmac100]

lest i forget,the maintenance program i have while being time-based involves the inspection and replacement of components after some hours of running.
one thing that i will need is the components that are affected by variations in certain parameters and the effect of the malfunctioning of such components on the operation of the units.
thanks again

 

[vmac100]

hi Romefu12,
sorry that i have not been in touch since,i have been trying to select my data and plan what i should do.i also had to enjoy the holidays,u know 'all work and no play sure would make Jack a dull and definitely unhappy chap'.hoping that you had a gr8 holiday too.
now,what i plan to do is this:i have got a time-based maintenance plan from Alstom categorised into class A,B and C.
class A- preventive maintenance due after 8000 run-hrs
class B- preventive maintenance due after 20000 run-hrs
class C- preventive maintenance due after 40000-50000 run-hrs
what i plan to implement is A,A,B,A,A,C.
however,we shall start counting the run-hrs from zero,after overhaul.the cost of the tools and the manpower required i have ascertained for each stage.
but what i want to add to this maintenance program is condition monitoring.i will propose the recording of the following operating data:
1.ambient temperature.
2.compressor discharge temperature.
3.exhaust temperature.
4.lube oil supply temperature.
5.lube oil return temperature.
6.inlet air pressure(downstream of filters).
7.compressor discharge pressure.
8.exhaust pressure.
9.fuel pressure(before and after start).
10.lube oil pressure(before and after filter).
11.fuel consumption.
12.lube oil consumption.
13.governor rack or metering position.
14.lube oil analysis results especially for trace metals.
i plan on proposing the collection of these data every 100hrs under the same load and operating conditions starting from when the turbines start running after overhaul.i the will plot the same data against the engine run-hrs.
now what would happen is that if there is any changes in the opration data/operating behavour over long periods then the next time-based maintenance activity,be it A,B, or C,would be carried out.the operating data would the be taken again for another 1000hrs and if there are still deviations,we apply the next maintenance activity.e.t.c.
finally,apart from the checks involved in those class of maintenance activities,i also would propose some extra checks that would be carried out on 1hr,8hrs,weekly,250hrs,500hrs,1000hrs,2000hrs,4000hrs and 8000hrs depending on whether the engine is on operating or standby mode.
i also would go to the plant and carry out a study on the failure mode and frequency of the components of the turbines and try to analyse the trend that is component in terms of component failure and how long it works before it fails.once i can get the average service time,i would propose that such components be changed before that time is reached to mitigate against their failure which can affect the other components.
as for the burner,i still have not ascertained the real problem.but i suspect that the diesel fuel quality is low.a purifier unit i will suggest to be used.
please any suggestions would be welcome,whether as regards the checks,the components i should watch out for,the effect on the failure of such components on the other components and the turbines,e.t.c. would be very appreciated.
thanks alot.

 

[Romefu12]

VMAC 100
Thank you for asking, yes I did have a great holiday, although busy. It’s good to get back to the grind. My own schedule is looking busier every day with upcoming projects.

I have a few questions before I begin to address each of your points. First,

1. Do you have an experienced Component Level Assessment Engineer to perform the A, B, and C inspections on the engine components or will you use the manufacturers representative?
2. Will you be using the time-based replacement of components without the use of an experienced component Assessment Engineer to perform engine performance analysis and root cause analysis?
3. What is the best time of year for you to perform your required maintenance? Spring, Summer, Fall, Winter?
4. Will you want to stagger the maintenance outages on these engines (usually a good idea depending on operational schedule) or perform all outages in series?
5. What are the normal operating modes of these engines? Base load continuous? Peak load? Backup / Fast Response?
6. Are all of these engines required to run all the time as primary power sources? Are any used as backup engines?
7. Where are these engines operated? (Geographical location, average climate, weather etc.)

I ask these questions because there are 2 practical ways you can go with this maintenance schedule depending on operating mode.

1. You can become a parts changer at regular intervals as specified by the manufacturer which can be much simplified (if setup and controlled properly) when compared to a completely performance based versus cost benefit schedule. This approach will make your parts, tooling and labor costs predictable but could be rather costly when replacing parts which are fully capable of running until the next scheduled outage or if your required maintenance schedule runs into your most critical operational periods (if any?). Remember we are talking about reliability and dependability therefore accurate scheduling and outage execution is the primary focus.

2. You can setup a condition based maintenance schedule with some of the parts changer schedule included provided you focus very closely on component condition, life cycle, and operating schedule. This scientific approach is what finely tunes the cost benefit of preventive and predictive maintenance on any equipment. This scientific approach is also what all power plant owners aspire to achieve because it provides the greatest of all worlds including the efficiency, predictability, greatest cost benefit, reliability, and availability. Let us not kid ourselves, it is these numbers that pay out your yearly pay raise and bonuses. When trying to schedule A, B, and C inspections based solely on operating hours without component assessment you will find that your operating hour scheduled maintenance inspections will eventually run into your most critical demand periods (if any?) as stated above. This is the not the time you will want to perform these outages and definitely not the time you need an engine component failure. You will also find that the various engine components will have very different lifetime cycles that will help to confuse the scheduling of parts replacement even further. Thus the need for experienced component condition assessment.

Now to address each of your points:

“What I plan to implement is A,A,B,A,A,C.”

This is a good plan if starting with new engines or starting from scratch with engines that need a lot of work. However this could be costly to perform a complete rebuild of numerous engines where a complete rebuild is not necessary. Thus an experienced component assessment would be more cost efficient for getting your maintenance program started down the right path.

“However, we shall start counting the run-hrs from zero, after overhaul. The cost of the tools and the manpower required I have ascertained for each stage.”

Same answer as above. This is a good plan if starting with new engines or starting from scratch with engines that need a lot of work. However this could be costly to perform a complete rebuild of numerous engines where a complete rebuild is not necessary. Thus an experienced component assessment would be more cost efficient for getting your maintenance program started down the right path. Knowing the cost is great however costs will change. Further, performing complete outages on engines that run the same amount of operating hours at the same time will virtually guarantee that all engine component life cycles will expire at the same time causing large loses in efficiency, reliability, and availability not to mention multiplied costs in labor, tools, and parts at the same intervals.

“What I want to add to this maintenance program is condition monitoring. I will propose the recording of the following operating data:”
1.ambient temperature.
2.compressor discharge temperature.
3.exhaust temperature.
4.lube oil supply temperature.
5.lube oil return temperature.
6.inlet air pressure (downstream of filters).
7.compressor discharge pressure.
8.exhaust pressure.
9.fuel pressure (before and after start).
10.lube oil pressure (before and after filter).
11.fuel consumption.
12.lube oil consumption.
13.governor rack or metering position.
14.lube oil analysis results especially for trace metals.

All of these are very good data points and can be used to trend numerous engine component life cycle predictions and optimum operating parameters when trended against the corresponding data.

“I plan on proposing the collection of these data every 100hrs under the same load and operating conditions starting from when the turbines start running after overhaul. I the will plot the same data against the engine run-hrs”

I will assume you mean complete overhaul. Data collection frequency is good however actual run hours will not directly correlate with EOH (Equivalent Operating Hours) therefore a calculation of predicted life cycle will be required for engine components. Unless you replace everything internally I would not start all parts at zero. I would keep all of the old data and make an educated hypothesis as to remaining life cycle in some components. Thus an experienced component assessment would be beneficial.

“Now what would happen is that if there is any changes in the operation data/operating behavior over long periods then the next time-based maintenance activity, be it A, B, or C, would be carried out. The operating data would the be taken again for another 1000hrs and if there are still deviations, we apply the next maintenance activity etc.”
I would not necessarily agree with this approach primarily because it sounds like you are becoming a parts changer. This method would also allow you to miss the root cause of problems that could be simply predicted, repaired, or eliminated on your normal outage schedule. The focus here would be trending and paying attention to all parameters. Experience with engines can really pay off here in predicting failures and conditions not specifically addressed in you’re A, B, and C inspection criteria.

“Finally, apart from the checks involved in those class of maintenance activities also would propose some extra checks that would be carried out on1hr, 8hrs, weekly, 250hrs, 500hrs, 1000hrs, 2000hrs, 4000hrs and 8000hrs depending on whether the engine is on operating or standby mode.”

What other specific checks would be performed?

“I also would go to the plant and carry out a study on the failure mode and frequency of the components of the turbines and try to analyze the trend that is component in terms of component failure and how long it works before it fails. Once I can get the average service time would propose that such components be changed before that time is reached to mitigate against their failure which can affect the other components.”

This data is extremely difficult to ascertain by anyone other than an Assessment Engineer, Component Life Cycle Engineer, or a Mechanical Engineer that specializes in Gas Turbine / Jet Engine technologies. I have numerous databases, spreadsheets, life cycle calculations, experience reports and analysis inspection procedures and approaches that address these very issues that I use every day. It has taken me 21 years of study in the manufacturing, installation, startup, commissioning, assessment and component design of these engines to understand as much as I do and I still learn something new on every engine. This will take up all of your free time and creative computer design capabilities to make sense of all the data you can collect. Component life cycle / failure prediction depend on the understanding of many factors including different types of corrosion, creep, fatigue, crack propagation, Thermal Barrier Coatings and spallation, Low Cycle Fatigue, High Cycle Fatigue, Temperature Transients, Component Cooling schemes, materials etc.

“As for the burner, I still have not ascertained the real problem. But I suspect that the diesel fuel quality is low. A purifier unit I will suggest to be used.”

A fuel purifier will be very beneficial if your storage tanks are clean, the purifier is used regularly, the fuel filters are maintained properly, and you have your fuel quality tested regularly. A couple of cents per gallon more for a better quality of fuel can save much. The burner problem could also be a simple problem of fuel valve timing, operating procedures or improper set-points in the turbine control. Burner fouling can also be a fuel rich operating condition. Your trend data should help you with this.

As a final input it would probably behoove you to include in your trending data the addition of a few data measurement points if possible:
1. Weather data including temperature, relative humidity at the times of your data collection.
2. Fuel regulator and / or fuel valve response times to loading /unloading inputs.
3. Governor position and response times versus input and weather data.
4. Starting, Loading, Unloading, and Shutdown cycles and times versus fuel valve/regulator position and timing.
5. Vibration monitoring and trending analysis.

Sorry for being so lengthy with a response but this is my job every day. Sounds like you have quite a project to keep you busy. I hope all of this helps.

Good luck

Romefu12

 

[vmac100]

Romefu12,
your contributions have been very helpful,thanks for the time u are spending.i have to say that i never knew that this would be this complex,when i took it up.one really needs a whole lot of experience to to tackle a project like this which i don't have.but i am not going to give up,it is turning out to be very interesting and exciting by the day and my boss seems to enjoy seeing me sweat it out,like i enjoy too.nothing like being in the middle of all the action.hahahahahahah!
now for the information you wanted,
1.i am ready to spend my spare time studying as much data as i can lay my hands on.
2.we have experienced staff on the ground but i still think formal training has to be done,as most of them got/are getting their experience on-the-job,which i think should be supplemented by proper training.however,such component inspection as they cannot carry-out will be given to the manufacturer's rep to do.
3.the replacement of the components will not be just time-based alone.i would propose that the condition of the component should be given priority over the time that it has run.so even if the component life-cycle has expired,if it is ascertained that it can still function till the next time-based general overhaul/component replacement,then it should continue to be used.certainly,the root cause of component/system failure would be asertained and cleared before the component is replaced.
4.the turbines are baseload continuous.one is going thru overhaul,one is to be replaced with a refurbished one, and two are running.IWANT MORE POWER FROM ALL OF THEM,ATLEAST 2.8MW instead of the present 2.3MW.
5.yes,i plan to stagger the maintenance outages.the plan will be implemented when we have atleast three of the turbines in good condition.
6.all the units are required to run all the time as primary power sources.like i said before,the actual plan is to operate based on an N+1 philosophy,but due to the fact that we don't have the turbines in good condition(which i still believe is due to poor maintenance culture),we operate an N+0 philosophy.
7.the geographical locaton is Nigeria(sorry that i did not mention it earlier,i am an HV Power/Project Engineer with the Corporate Utilities Department of The Shell Petroleum Development Company of Nigeria Ltd.i would give you my name outside the forum) and the environment is tropical.
8.the suggestio that all parts should not be started at zero is a very correct one.i will keep that in mind.
i really have to thank you for your help,i will continue sourcing for information and hope that you will be of further assistance.
thanks.
n.b:sorry that this was lenghthy too.

 

[vmac100]

Romefu12,
i would like you to look again at those parameters that i listed and suggest which ones you think are the critical ones.
i think that when there are large deviations in the measured values of some parameters from the reference/base value(like the value one gets after say 100 run-hrs after overhaul)then there is a problem with certain components.what do you think?

 

[vmac100]

Romefu12,
can you suggest some components and the related parameters i should watch out for?
thanks

 

[Romefu12]

vmac100
Are these particular engines your only primary responsibility or are you responsible for an entire fleet of engines?

We would probably need to communicate outside of the forum. I say this primarily because our discussions are getting deep into troubleshooting, predictive, and preventive procedures that are primarily OEM information. I could teach a complete degree program in gas turbines here if I continue. What you are asking me to answer gets even more involved and more lengthy to explain. Pictures and graphs would then be necessary and some life cycle calculation programs wouldn't hurt however I would not reveal them here in the forum.

Of course if you would like to hire me full time, I could be persuaded by a large company.

I will continue to help where I can but I would need time to put together a few things to model.

Romefu12
Romefu12@comcast.net