Emergency Diesel generator 1

[teke]

Please tell me CAN I NOT KEEP A EMERGENCY DIESLE GENERATOR IDLE FOR PROLONGED PERIOD? It is connected to a switchgear bus with utility connection on the other main. Utility supply rarely fails, so the emergency diesel gen is hardly used.

 

[catserveng]

When you say 'idle" I assume you mean not running at all. Or are you asking if can idle the engine for extended periods of time?

Generator sets used in standby applications need to be exercised regularly to assure they will operate properly when actually needed. If you don't regularly test run the engine how will you know it will crank, start, run and produce needed voltage and frequency when really needed?

Depending on your facility and local regulations, you may need to test as often as once a week. In general, most diesel engine driven generator set manufacturers recommend at least a monthly start and run test and at a minimum a full load test once a year for 2 to 4 hours. Your insurance carrier may also have requirements you need to meet.

You also need to make sure the engine has a jacket water heater and the generator has a space heater. I recently found a number of standby units where the heaters have been shutoff as an energy saving measure, not a very good idea!

Hope that helps,

Mike L.

 

[teke]

Thanks for your quick comments.

When you commented about monthly start and test, it appears to me it is on no-load, so it will be probably a run duration of 5-10 minutes typically?

 

[racobb]

Do not do your tests at no load, they need to be at least at 20 to 30% load. I will let catserveng tell you why.

Alan

 

[rbulsara]

Look up wet stacking. For example, see this link:

http://www.kraftpower.com/pdfs/KPC_IS_16_Wet_Stacking.pdf

Rafiq Bulsara

http://www.srengineersct.com

 

[catserveng]

Generally we like to any test run done at load, however in practise it is usually VERY difficult to get an end user to either pay the expense of a portable or permanent load bank or use their facility loads for testing.

Hospitals are the worst in my area, getting them do do a loaded run is like pulling teeth, and finding money and space to install a load bank just doesn't happen.

If you are going to do a no load start and run test, limit it to 10 minutes, let the engine start, come up to speed, monitor voltage, frequency, oil pressure. Make sure the fuel system, cooling system and other support systems are working properly, check the engine for leaks. Stop the unit, make sure the battery charger is working properly and that all controls are reset so that the unit functions properly in AUTO.

As a general rule, for every hour on a standby engine you run unloaded, you should run at least once a year the engine should run at or above 50% load for an hour.

The majority of customers I used to support when I worked for a dealer did weekly or monthly start and run tests, and an annual load bank test for 4 hours.

If you have any design or operational influence, having some kind of load to test downstream of the ATS is preferable, because it also needs to be tested. But few customers, except high end data centers, will put in the appropriate bypass cicuits or design for regular testing.

On large telecom customer we used to have did weekly start and run tests, quarterly site load tests thru their ATS, and an annual load bank test during the hottest summer months. Overall after about 20 years we found this customer had one of the lowest numbers of problems with his standby systems. Of course a lot of other things went in to assuring system performance as well.

No load start and run test only give some assurance the unit will start and go to rated speed and voltage when called. These types of tests do not fully prove the ability of the support systems, such as fuel supply, lube and cooling, to properly function when the called for an actual emergency.

The earthquake in Hawaii and subsequent power outage on Oahu a few years ago is a great example, about 65% of the standby generators "failed to operate as expected". Many failed to start at all when called, but most failed and shutdown with the first hour due to mainly fuel system and cooling system problems. Virtually every unit that failed had some form of "regular maintenance and testing" performed.

Since I don't know what type of facility your supporting on can only offer fairly general advice, but your generator set vendor is likely your first best source of information for you unit and application.

Hope that helps,

Mike L.

 

[davidbeach]

The best is to have the gear that allows paralleling the generator with the utility. Then the test run is just parallel in and ramp up the load.

 

[waross]

But don't tell the utility. And I know of at least one utility who's default meter setting will bill you for the power that you give to the grid just as if you had used it.
But I agree with David. If you can get away with it, the grid will make a wonderful load bank.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[davidbeach]

What is there to "get away" with? If you're going to parallel with the utility they need to (absolutely must) be involved in the design, but it isn't that difficult. Most hospitals and other facilities with emergency generators don't have enough generation to offset all of their load anyway.

 

[mmt019]

I am assuming "idle" means that the generating is not running but is available to be started for backup power. For general emergency backup electrical power purposes, once a week "no load" generator exercising for 1/2 hour is acceptable with at least once a year 100% load testing (for 3-4 hours). Paralleling depends on the application, load testing reduces wetstacking.

 

[catserveng]

In the area I do most of my work, southern California, I have about 6 facilities that have large standby systems and the appropriate switchgear and controls to parallel with the grid. The main design intent was to be able to parallel and soft transition to and from the grid in the event of a possible power disruption. The side benefit was in all cases after approval of the utilites involved was that we could parallel and load test units as long as we didn't export power back into the grid. These sites have what is likely our most expensive switchgear and control systems for standby operation, but the benefits to the customers made them well worth it.

We also have several units all over soCal that are involved in utility demand response programs. The utility working with a contractor has retrofitted many sites that have larger standby power systems with parallel switchgear and controls and required exhaust aftertreatment. The utility requires regular testing and pays for maintenance and costs associated with demand response when needed.

If the work is well coordinated and done, and the right testing and maintenance programs are performed, I have found these to be very beneficial to both the utilites and the equipment owners. Not sure how many of these type programs are operating but if an end used has a standby system he may ask his utility to see if his equipment and site are a likely candidate.

Many ways to achieve proper care and assurance of your standby power system, but a program does need to be in place and followed to assure your unit performs as expected in the event of a power failure.

Hope that helps,

Mike L.

 

[TheBlacksmith]

Sorry to be crass, but starting and running your emergency generator at no load is like lighting all your matches before a camping trip to make sure they work.

 

[waross]

Well, I have seen too many standby sets with speed/trip issues to agree. Just because a generator started and carried the load when the grid failed does not guarantee that it will start the next time.
When a set is carrying a heavy load and the load abruptly transfers back to the utility, the load dump results in an over-speed transient. I have seen a number of occasions where this has resulted in an un-noticed over-speed trip which was not discovered until the next power failure when the set did not start because the overspeed trip had not been reset.
Any supervised start gives you an opportunity to evaluate possible problems developing with batteries, starter(s), and general engine health.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[catserveng]

Over the years of working on many standby systems, I do find good reasons to do a regular no load start and run test if test running under load is not an option.

As Bill points out, many units can suffer problems at the end of an actual outage run. A regular no load test run does test the engine batteries, cranking system, governor and voltage controls, and other system functions. It does not assure the unit will carry site load, but it does give a reasonable assurance the unit will start and run.

It should be considered PART of an overall system maintenance program. It is never satisfactory as the only means of testing, and is way better than not operating and engine at all for months on end. At least my two cents worth.

Mike L.

 

[TheBlacksmith]

Like I said I was crass. I agree that any start and run is better than none. However, in my experience (Navy ships and predominantly 2 stroke cycle), I have seen too much port clogging and cylinder liner deterioration from prolonged light loading.