Engineer types: How can I load a big Cummins generator?

I need some engineering input.

I'm doing a test of some diesel fuel products. We have a need to test multiple products over an extended period head-to-head.

The target audience for the product is over the road trucks. By far the most common engine is the Cummins 12L.

I am trying to set up a test that will see how different products perform in a similar engine.

Architecturally, the 12L is similar to the B-series 5.9L diesels in Pickups, and the 3.9 in generators and gensets. As long as it is a direct injection motor that runs at about 1800 rpm.

The problem with using a truck for the tests is designing a test protocol that will exclude the influence of the engine management system. We could put an engine on a stand, but the computer could make all sorts of adjustments without us knowing it (Oh look- an improvement! No, wait...is it running in limp-home mode??)

So we are thinking we can get a generator with a 3.9L and have a better test procedure, because of the simpler control systems. Maybe even find one with completely analog controls.

So, once we have the unit, we need to run it off and on for several weeks or months. But idle doesn't mean anything. It MUST be under load.

What is the best way I can put a constant and consistent load on a generator so it will remain that way for an extended period? I'm thinking a 50kw unit, so don't say toaster oven.

We thought of running the plant off it, but that would be an inconsistent load. I'd love to backfeed the power grid, but the red tape involved is prohibitive. I need this thing running pronto.

Thoughts? Big heaters? Other?

Other ideas for setting up a test?

Phone Jack Roush and ask him. He sells electricity to the city from his dyno cells....

Can't backfeed the grid. Local and state redtape would take months to get it up and running.

Probably a stupid idea, but I'll throw it out there anyway. How much power do the light in the plant draw? Could you isolate a circuit that would be consistent load hook it up to the generator and just leave what ever that is running?

A quick calculation seems to show that swimming pool pumps are about 1kW/hp. If you have a place for a reasonably sized tank of water some kind of water pump might be a good load for the generator.

MG_Bryan wrote: Probably a stupid idea, but I'll throw it out there anyway. How much power do the light in the plant draw? Could you isolate a circuit that would be consistent load hook it up to the generator and just leave what ever that is running?

Plant's not big enough.

It's a real shame you can't feed the grid.

Our dyno uses hydraulic pumps for a load. The fluid is pumped through a couple of big heat exchangers to keep the temperature under control, and there are two big 40hp fans driven by the hydraulic pressure to pull air across the exchangers. Pretty clever, actually. Something like that might work, although the swimming pool pumps are simpler.

How about offering a free electric car recharging station?

HoustonNW wrote: A quick calculation seems to show that swimming pool pumps are about 1kW/hp. If you have a place for a reasonably sized tank of water some kind of water pump might be a good load for the generator.

Hmmm... the beginning of an idea. Large tanks (up to 10,000 gal) are available. Maybe we could set up a loop pumping water from one tank to another. Now where am I going to get 40 or 50 swimming pool pumps?

Keith wrote: It's a real shame you can't feed the grid. Our dyno uses hydraulic pumps for a load. The fluid is pumped through a couple of big heat exchangers to keep the temperature under control, and there are two big 40hp fans driven by the hydraulic pressure to pull air across the exchangers. Pretty clever, actually. Something like that might work, although the swimming pool pumps are simpler. How about offering a free electric car recharging station?

Electric car... in Albany, GA??

How big is your dyno? Sounds like it is designed for trucks or jumbo jets...

How much power do those heat exchangers absorb? How big are they?

your 10k tank, large centrifugal water pump, de-aerator, discharge valve to load up ?

I just got home from witnessing a generator load bank test. I work for an engineering firm that does a ton of this work. If you need a consistant load that you can repeat multiple times I'm afraid you're looking at a true load bank. Basically they're just a big series of heating elements that can be adjusted to the desired load. They're available for rent but it won't be cheap.

Any kind of plant load is going to be too variable. The pump idea isn't bad but pumps tend to load up and surge as you get cavitation. It's hard enough to get meaningful results without having your load jump all over the place.

Just spitballing here but large space heaters may provide a more grassroots approach. I'm thinking the big 5kw ceiling hung type units. Ovens may work too. Let me sleep on it and see if I can come up with something better/cheaper.

SVreX wrote:

Keith wrote: It's a real shame you can't feed the grid. Our dyno uses hydraulic pumps for a load. The fluid is pumped through a couple of big heat exchangers to keep the temperature under control, and there are two big 40hp fans driven by the hydraulic pressure to pull air across the exchangers. Pretty clever, actually. Something like that might work, although the swimming pool pumps are simpler. How about offering a free electric car recharging station?

Electric car... in Albany, GA?? How big is your dyno? Sounds like it is designed for trucks or jumbo jets... How much power do those heat exchangers absorb? How big are they?

Hey, if you're generating 400 hp, that energy has to go somewhere. The exchangers don't absorb energy per se, they dissipate the heat generated by the pumps. The cooler units can be dragged across the floor by one person with some effort, but it's easier to move one with a forklift. Still, the whole rig will fit in the bed of a pickup. In order to provide a consistent and controllable load, the output of the pumps is restricted with an computer-controlled valve running in closed loop. Strain gauges measure the actual torque at each hub. It's pretty cool stuff.

http://flyinmiata.com/tech/rototest.php

Can you find a local business to sell/give the power from a generator to and house the generators on site? Maybe one that's running something pretty constant, like a pumping station or something?

Streetwiseguy wrote: Phone Jack Roush and ask him. He sells electricity to the city from his dyno cells....

+1.

And not HOW to backfeed the power, but to have Roush Industries do the testing for you. It will probably be cheaper in the long run, and odds are that they do have access to the computer system so that you can monitor the controllers effect on your whatever.

I personally don't know anyone in the diesel industry, but am pretty sure that you can find a sub-contrator for cummings who can do the dyno work for you. And include all correct measurements so that your results are good.

Home made drum dyno.

55g drum filled with concrete. A big belt, some bearings & pillow blocks, a tachometer driven from a toothed timing gear and hall effect sensor... calc the load, and keep the RPM constant.

Since you don't particularly care about exact HP/Torque graphs across the range of engine RPM and as long as it is kept consistent across tests... that should work just fine.

On a side note, I'm not sure exactly what you are testing- but I would run full ECU control, and monitor the system if there are any effects.

Assuming that the product you are testing is intended for a diesel, on road, application.

The reason is that IF there is some kind of adjustment, you should know before it happens- nothing worse than to sell a product on it's benefits only to find that the side effects are worse than the improvements. You can monitor how the engine is running, and record it. A stationary dyno for a pick up would work quite well for that.

Again, assuming that this is for a real product, I would very much avoid backard or DIY solutions since they will bring as many or more questions to the data, let alone the development of that, to not use your time or money wisely. If this is a real, on road, product, SAE procedures are a good idea to follow, as well.

Which will generally bring you back to facilities that are already capable of doing the testing. Unless, in a long term plan, it's also part of the plan (which also means returning power to the grid as a long term solution).

Seriously for 50kw... 100x 500watt light bulbs (or 200 of the 250 watt bulbs - easily found at your local home depot). If you need to make a minor adjustment to load, you just remove/add a few bulbs. Or you could get fancy and add switches for each row of 20 or so I imagine it would take someone about 8 hours to build it all onto a few sheets of plywood?

(and for those referencing a load bank... ^that's all a primitive load bank really is, but with more precise resistors than light builbs)

Wayslow wrote: I just got home from witnessing a generator load bank test. I work for an engineering firm that does a ton of this work. If you need a consistant load that you can repeat multiple times I'm afraid you're looking at a true load bank. Basically they're just a big series of heating elements that can be adjusted to the desired load. They're available for rent but it won't be cheap.

This is probably the only real answer. I'm in the same kind of engineering, we deal with generators all the time and it's typical that before delivery the generator is tested with a load bank.

It turns out there's a Wikipedia entry for load banks: http://en.wikipedia.org/wiki/Load_bank

Check up on the EPA NESHAP for diesel engines. As well test cells, and electric generators. You're going to make yourself run afoul of a few of them doing what you propose.

I'll disagree with you about a computer changing things on the test cell. Quite the opposite, it will hold things exactly the same unless you tell it to do otherwise. That's what test cells are all about, control.

If you're not worried about energy recovery, waste heat generated electrically is very easy to regulate and hold steady. Water pumping as well, but you can't adjust it as consistently as you can electrical waste heat.

Dynapack dynos drive an oil pump through a restrictor valve and use an oil to water heat exchanger to cool the system. Might be able to fab something up with an industrial hydraulic pump.

Pumping water in a loop consistently is all about head pressure, or the height of the water column. Put your tank of water up high, allow gravity to feed the pump, and the pump to lift the water back to the tank. Some pretty basic calculations will allow you to determine the height and size of pipe you will need. Adding a restriction to the inlet to the tank can also be used to increase head pressure required.

wha sort of dyno does cummins use.

In reply to Wayslow:

I agree the load bank is a great solution, and the only acceptable solution for the firm you work for. The cost of the testing we are doing would hit about $250K if we hired you to do it. I need to get the first phase done for more like 10% of that.

In reply to Osterkraut, Alfa, etc:

The idea of working with Rousch or any other company is pretty good. Just got to figure out how to maintain confidentiality with proprietary information.

Regarding Dynos:

We've already tried a real full sized over the road truck on a dyno. Doesn't work so good. The dyno is designed to provide a load that battles the energy generated by the vehicle. Essentially, there are 2 separate computers (truck and dyno) both of which are trying to optimize what they do (generate power and offer resistance). Neither is a steady flow, and therefore not optimized for measuring improvements to fuel burn.

Additionally, the dyno will overload with heart sink when the unit is run for an extended period.

ECU's or computer drive systems are compensating for the very stuff we are trying to measure.