What RPM's does it run at? And how many cylinders? Is it 4 stroke or 2? And what kind of failures are the belts experiencing? How are the pulleys mounted to drive from the engine? Are all the belts you have tried from the same manufacture? Is this gen set from a large manufacture or from a small out fit that just assembles engines and generators? Because most large reputable outfits will test their sets for thousands of hours to prove they will last.

At first guess, a 2MW genset using a single engine would be up in the range of EMD power.

Belt drive to fans with these engines can be tricky.  Most  locomotives I've seen use electric fan drive (the fans are remote from the engine, and operate in a different plane).  One locomotive (a Baldwin AS-16 with original DeLaVergne engine) had a strikingly large number of V-belts between crank pulley and water pump -- I didn't count the exact number but it was quite a few.

Be sure you've debugged everything mechanical before going back to specs... no insult intended to professional capabilities ;-}

Is the water pump on this genset driven by the fan's belt?  

For the belts to be 'failing' at 150 hours is unusual.  Are these standard V-belts or poly-V?  What method are you using to tension them?  I presume you have checked bearings, resistances, blade-to-shroud clearances, etc. and have confirmed that nothing obvious is contributing to load on the belts... Be aware that the failure might only show up at higher speed (e.g. fan nutation from a bad bearing) and, as with some driveshaft-bearing failures, might not expose itself to simple tests (like pulling on the fan to detect play).  I tentatively rule out a fan-clutch problem simply because I can't see a way that a bad clutch would create both a "cooling problem" and a belt-wear problem at the same time -- but check it out anyway if you have one.

Be sure to check the internal profile of the pulley grooves.  Wear here can eat belts fast, and 'reading between the lines' this genset may have eaten enough belts to cause pulley wear even if that weren't original contributing cause.

Potentially irrelevant anecdote:  I was salmon-fishing, and walked in on some of the staff who couldn't keep the camp generator from cutting off.  The unit was diesel-powered (about the size of a 6-71 IIRC although I don't remember the exact engine manufacturer) and the high-temp engine cutoff was pulling in -- staff was in the process of blocking this (!!!) out of frustration in finding a cause.  Coolant was correct, fan drive intact, no visible obstructions in the cooling air intake or exhaust.  Turned out there was a piece of screen that slipped across the radiator intake, not well-documented on the genset itself, and this was almost completely plugged with mosquitoes.  Moral: be sure your core is actually seeing the right flow of cooling air, rather than just being sized correctly...

Answering the question you asked:

1.  A request to a company for technical measurements of a piece of equipment you own should be met with courtesy.  I specifically define that to include releasing technical information about size and capacity of cooling components.  Whether or not "other customers" seldom "need" such information is completely irrelevant to YOU as a customer who does need such information.

2.  There are very good texts on heat exchangers -- at least one of which (from MIT as I recall) has been published to the Web as a public service.  Any one of these should contain both formulas and knowledge to determine heat rejection and airflow for a radiator core.  You will need to know your thermostat setting and type (e.g., does it open completely when actuated, or does it throttle to maintain a particular temperature or range of temperatures across it)

3.  Likewise for fan curves: fan manufacturers publish design guides that contain the formulas, and some of the discussions to use them best.  I don't have references at hand, but have seen a number of these, including some historical ones from the turn of the century.  Another potential reference source might be HVAC texts (rather than engine-related ones).  You may not find as much sophisticated discussion of 'typical' engine fan types, as much of the 'action' for high-efficiency fans in HVAC involves large air masses at relatively low speed.  On the other hand, some applications (such as ceiling fans for circulation) are now involving careful aerodynamics in blade design and optimal speed-range determination.

Hope some of this, at least, is helpful...

RME

xr7755,
To answer some of your questions:
16 cylinder 4 stroke engine RPM=1800, Fan RPM=900. The belts are synchronous (cogged
). The belts snap suddenly.  The genset was built by a major manufacturer. The engine/alternator are standard pieces for this manufacturer, but the cooling system is a custom design. The engine crankshaft sprocket drives the fan sprocket via the cogged belt. There is one inside idler, used for tensioning, on the drive side of the belt. The fan and idler sprockets are mounted to a steel frame bolted to the floor of a shipping container. The radiator is bolted to this frame. The engine is mounted to its own steel “skid”. These two structures share little connection.  No intentional provision for alignment was made by the packager. We must unbolt the radiator and sprocket structure and coerce them into alignment.

You haven’t lived until you have witnessed an EMD take a 2MW block load. Hearing and feeling the turbo come off the gear train and spool up is almost better than ...I digress.

overmod,
    This is not an EMD engine. We use remote radiators with electric motors to cool our EMD power plants. I have checked the things you mentioned and they seem to be ok. Thank you for your input!

I suspect this isn't the case, but if the gen-sets are of either Cummins or Caterpillar origin the cooling requirements and radiator air flow can be found at their respective websites.

  Just a thought, could it be that the idler pully sometimes hangs up (over tensioning the belt) thus spap-ola just a thought.

From my understanding of your set-up the engine and alternator are soft mounted to the skid/baseframe (rubber or spring mounts usually) and the radiator, fan and idler are mounted separately to a steel frame which is bolted directly to the container floor.

This means that the engine is free to vibrate but the fan which it is driving is not. This could be where your problem lies.

Have you considered torsional harmonics of the engine crankshaft?  I have no idea if this is your problem or not but I've understand that it can be an issue with belt driven accessories.

If I understand correctly, the belts are cogged?
You mean the same as a standard timing belt?
As stroker six mentioned, that was why I asked the questions. Was thinking torsional loads. And with a cogged belt that could be that case. You may want to go to multiple vee belts. They allow for slippage.

If the engine/alternator are "standard product" from the genset manufacturer then it is very probable that there should be no problem from torsional vibrations of the cranktrain. I can think of 2 possibilities however that might lead to torsional problems with this "special product":
1. The fan is pulling a lot more load than the "standard" fan arrangement, which will change the properties of the torsional system.
2. The torsional damper is usually built into the crankshaft pulley system and maybe this has been adversely affected by the new fan drive system.

Another point to think about.

If the unit gets started often and if the generator starts running up to operational rpm directly, as some do, then there might be a lot more torque than the belts are calculated for during the startup. In this case maybe you should try V belts instead allowing for some slippage.

V belts it is! This seems to be the consensus all around. Conversations with all the component manufacturers involved tend to lead toward the fact that cogged belts are not the best means of power transmission for this application. I am doing some reading on vibration to learn more. From what I understand so far the dynamics of this system are quite interesting and not necessarily intuitive upon first look. Any reference recommendations? Thanks to all.

Gentech,

Looking at some catalogs on belt drives I wouldn't jump to conclusions based on the above at this stage. You may just have the wrong belt spec. with regards to type of reinforcement. I would contact people like gates.com and let them ask the right questions and see what they can come up with. If you see what is available and the multitude of options there must be a way, unless the cog wheels are just too narrow for the application. It would look pretty stupid if you went for V belts to discover later that cog belts was the right choice afterall but you only used belts with the wrong spec.