Hello all,

I am new to this forum. I'm a surgeon who ages ago was much into electronics, minored in EE. Still quite handy, but my circuit design knowledge base has decayed significantly. I'm trying to design/build a dynamic resistive load bank for a 5k Diesel Generator set. I want to keep the two hot legs of the 240 line balanced, and the whole thing running at no less than 60% load. I am interested in building two paired load banks that do the following:

1. Sense the current running through a generator output source line
2. When the current is less than 15 amps (high side), the load bank is activated and adds load to that line. When current is over 15 amps, the load bank is off.
3. The amount of load added is equal to 15 amps minus the actual loading on the line This ensures that no matter what is one in the house, the generator is always loaded at a minimum of 60%, and it stays happy.

I know one could easily do this manually with light bulbs, or space heaters, and while I'm open to using them for loading (as opposed to very pricey grid resistors), I'm interested in solutions that allow continuous variability in loading, and the sensing component so that when the family turns off a television, I don't have to run to the garage and turn on a light bulb.

Any guidance would be most appreciated. One thing I'm not considering is some scheme to raise and lower rods in a bucket of salt water - sure, it would work, but that's not the direction I'd like to take this. thanks,

Nelson

 

This request is so strange that I have to do something I rarely do...ask why.

My first instinct is, "Fix the generator" so it runs like all the others, adjusting itself to the load as required. That is pretty presumtious, so I ask, why do you want a self regulating load on a machine that is supposed to regulate itself?

 

I have a portable diesel generator. Regardless of what the generator is capable of putting out, current flows based on the connected load. If my house is drawing 2 kwH, which is what I average, at any one given time, only 16 amps are flowing out of the generator, which would be 40% load. The lower the load, the less demand on the engine, and diesels are high compression engines that have some serious problems when they run light loads for long periods of time.

Two complications set in - glazing, and wet stacking. Both of these can cause permanent engine damage. Diesel generators like to run loaded - it is unclear what is the optimal loading to eliminate risk of glazing and wet stacking, while maximize fuel efficiency, it varies by generator. Most agree that the generator is happy with a load over 50%, and probably closer to 70% is optimal.

Big gen sets do this with load banking, commercially available devices to ensure that when your running a 100k generator, or a megawatt, on a light load, you don't destroy it. gas generators don't care in the slightest. Then again, a gas generator may last 500 to 1000 hours. A diesel can last 50,000 and does so running on half the fuel. There are adavantages to diesel. Plus, during my recent 2 week outage from Sandy, diesel was ubiquitous, but gas lines were worse (in my area) than 1974.

The reason for two paired devices is for load balancing between the two poles of the generator - so you don't have 2 amps on one hot line and 18 amps on the other. That's not good either, although I am unclear on what the long term sequelae would be to that arrangement.

Most folks say that with smaller generators, this whole deal is less of a concern. That's why i'm not shelling out $3000 for something like this. However, I am interested in doing this. can anyone suggest ideas? I figure I need a normally closed current sensing relay to keep the thing on until the current is enough to shut it off. Then, you could use the output of the current-sensing toroid to feed into whatever circuit would vary the resistive load - resistive loads are common, cheap, but crank out heat. Still, may be the easiest way to go. How would one use the signal from a current sensing toroid to vary a resistive load?

Nelson

 

That was really informative! (Very few electronics nerds know that much about diesel engines.) The torroid of which you speak is called a current transformer. Very standard stuff. The obvious way to get rid of power is to turn it into heat. Any other method would be peculiar to the individual situation, like pumping water into a reservoir tank during low load periods or making ice to use for air conditioning later, when there is no surplus power to run the air conditioning compressor. Do you have any opportunistic loads?

I'll start working on a block diagram and tack it on to this later.

Meanwhile, other people have a good idea what you need and they will also answer.

 

If my house is drawing 2 kwH, which is what I average, at any one given time, only 16 amps are flowing out of the generator, which would be 40% load

Is this load balanced? I think if you put this load on one phase (if possible)it'll cause your generator to work harder to produce that power(2kwH) at 120 volts.
The remainder of the load would be on the other phase, to balance it when under a heavier use.

just my $.02

 

Yes, balancing is an issue, but again I am not sure what's the downside to running an unbalanced load over the long term - maybe there's assymetrical wear within the alternator?

To balance the loads, I figured I would have two of these devices, one attached to each hot leg of the 240 output. That way, regardless of which house circuits were on either leg, the 'adaptive load bank' would be able to both ensure optimal loading, as well as balance the loads across the generator's two poles.

-Nelson

 

I'm going with, "keep it balanced". A horsepower is a horsepower and a kilowatt is a kilowatt. Trying to pull the power out of one phase takes the same amount of power as if the load was balanced, except one winding of the alternator will be hotter than the other winding and one set of slip rings will be more worn than the other. You have to pay to rebuild the alternator when one set of slip rings is worn out. Better to use them equally and get twice the lifespan.

 

My home averages almost no electricity consumption. At night when electricity costs less we do laundry and cook a late dinner. All light bulbs are low current compact fluorescent. Our christmas tree has low current LEDs supplied for free from the electrical utility company.

 

First of all - the swift-e is a "test" load bank. It is not intended for long duration load dumping..
....
You have a novel idea - I really like it. Takes me back to Turky and the USAF Tropo Site Load Balancing Scenarios...... But dynamic load balancing is hard to do in "auto sensing" mode.
...
Still - it has been done, and I am sure the details are drifting around on the web someplace.
..
One Idea comes to mind (and you tripped this thought by your reference to the recent east coast hurricane disaster) - Electric Water Heaters (even electric boilers) make excellent long-term load banks. Combine this idea with the practice of solar space heating using stored hot water and you have a very doable project.
..
You see - We use hot water to take showers, but hot water is a really efficient (and can be reasonably stored in Insulated storage tanks). In the solar energy world the storage tanks fall under the heading of "thermal Mass Storage Vaults". A lot of energy is put into boiling water - but if you boil water in off-peak and low demand periods - you are in effect load banking. Boil enough water and store it efficiently and you have an off-peak load for the GenSet and hot water to spare. Couple this with radiant heat coils in the floors of the residence and you have creature comfort, hot water and a dynamically loaded GenSet (which is your ultimate goal - Maintaining 65% critical loads). In Turky we used three 150KW Generators to supply power to a site that had a minimal power requirement of around 120KW with peaks at round 260KW. Power pro load banked by using switch gear to parallel Two generators during high load periods, and shut down to one during light loads. We also boiled water when the loads dropped to less than 65% in the early morning hours when most of the site were asleep and office personnel and admin offices were idle.
....
A very simple sensor can be fabricated whereby the loads are sensed and should the load drop to a pre determined level - would turn on the immersion heaters and boil water. To make it truly dynamic you could have two sensory/loading circuits one for the A phase and one for the B phase (should you have a truly three phase GenSet - simply build a system that works on all three phases. The beauty of it is that you can go as far as scale the loads (immersion heaters) so that you can dump say 1 or 2KW, maybe 3KW, 4.5KW, 6KW, 7.5KW etc as needed to balance the loads across the active power legs. So should the A Leg of a 20KW Generator be seeing 4KW of load, while the Second power leg was seeing 10KW - you could scale 6KW f load to the A power leg, and then scale an additional 6KW + 6KW to attain 16KW loading across the two power legs - effectively balancing the power.
..
There are literally tons of current sensors out there and probably an infinite number of scenarios that would be suitable to do load sensing. The difficulty would be applying the logic to when and how much, then the ability to actively engage or disengage the loads to maintain a reasonably steady load rating.
...
I am going to watch this thread with anticipation. I am looking forward to the reply's from others.
..
At the same time I will be looking into the project to see if I can provide more technical references and ideas.
..
Right now I am tied up and cannot devote a lot of time to the research needed. But give me a week -
...
Dave
Phoenix, AZ

 

To start I'm no expert in these specific matters,just a geek who sees an interesting problem and has a spare two cents to toss into the ring.

Let me see if I understand this. Basic system monitors the load power, and turns on devices to keep the load above some minimum. And as the system is a 120/120 V system each line needs such a monitor.

Seems very doable, and probably can be done with mostly off the shelf parts, though one of those parts may be an old PC to act as the controller. The system breaks into 3 pieces: power monitor, controller, and load switches.

The Google PowerMeter used an Energy Detective (TED 5000) device to read the power. This device uses the power line to send data, and also has a remote unit to convert the power line signals to a wireless ethernet signal to get it into it's base unit, or a PC.

The load switches are just simple on/off devices. Using 100W bulbs you could string them together into 1, 2, 4, and 8 bulb strings which gives you a 0 to 1500 watt load, 100 watt resolution, for only 4 switches. There should be some sort of home automation device that could handle this.

The PC in the middle is the hardest part, as it would need some sort of custom software to receive power data and convert this to switch commands. It's not a very involved program, but it is a program and will take some programming smarts to put it together.

That's my spitball idea for the day.

 

Ok , so approaching from another angle. You want the generator to run full time on at least a 70% load, is that correct? And as other (read necessary) loads come in you want the load bank to drop out?
First thing that comes to mind in that scenario is to attach current transformers to the power lines AFTER the load bank. the output of the current transformers could turn off the load bank as the necessary load turns on.
The load bank, being before the current transformers, isn't seen by then and therefore not measured by them.
On another note, diesels idle all the time without doing any damage to themselves, can you relay the source of damage you spoke about?

 

A bit of further thinking results in a drawing.
Based on the idea that this machine does not have to respond to a decreased load instantly, I think an analog method would work...as in, "Doesn't need a computer".
Right now, I'm lackng the skills to design the increment/decrement circuit.

 

I'm going with, "keep it balanced". A horsepower is a horsepower and a kilowatt is a kilowatt. Trying to pull the power out of one phase takes the same amount of power as if the load was balanced, except one winding of the alternator will be hotter than the other winding and one set of slip rings will be more worn than the other. You have to pay to rebuild the alternator when one set of slip rings is worn out. Better to use them equally and get twice the lifespan.

That's assuming a brush alternator. Better alternators utilize brushless designs.

 

I'm trying to design/build a dynamic resistive load bank for a 5k Diesel Generator set.

Having spent a few years in the portable generator business, I am very curious as to the make and model of your generator and its engine.

 

I learn something almost every day on this site!

 

We're heading in the right direction. Focusing on a circuit to do this, consider this possible algorithm, and how one would design it:

1. Sense the current flowing in the high-side
2. Set the desired minimum load (range 10 to 15 amps) with some variable device (?potentiometer)
3. Compare the high-side current with set point, and calculate the difference
4. Based on this difference send signals to relays to turn on, or off, a sequence of light bulbs (if you used 6 halogens, 300W x 4, 200W x 1, 100W x 1, you could add additional load to the line in 100W increments and you could get every value from 100W to 1500W with the fewest # of bulbs, without exceeding 300W bulbs, to keep cost and heat down).
5. A small fan to blow over the bulbs while it is running, just to cool things a bit.

The crux of the design, circuit I'm thinking about has a comparator and then some way of turning the result into a logic circuit to decide what to turn on (or off).

Thoughts?

How would one design a circuit to do that? Each light bulb would presumably have it's own relay.

 

Yanmar YDG-5500. Picked it up a few weeks after Huricane Sandy. As I am not really using it, I should be testing it monthly and running it on load for at least a half hour. While there is much discussion on load banking in giant gen sets, there are no design reasons I can think of why the same principles should not apply to my 5K. I need the 5k because despite my average use of about 1.8KW (wintertime; no AC), I have two sumps and must be able to accomodate the inrush current on their 8A motors.

 

We're heading in the right direction. Focusing on a circuit to do this, consider this possible algorithm, and how one would design it:

1. Sense the current flowing in the high-side
2. Set the desired minimum load (range 10 to 15 amps) with some variable device (?potentiometer)
3. Compare the high-side current with set point, and calculate the difference
4. Based on this difference send signals to relays to turn on, or off, a sequence of light bulbs (if you used 6 halogens, 300W x 4, 200W x 1, 100W x 1, you could add additional load to the line in 100W increments and you could get every value from 100W to 1500W with the fewest # of bulbs, without exceeding 300W bulbs, to keep cost and heat down).
5. A small fan to blow over the bulbs while it is running, just to cool things a bit.

The crux of the design, circuit I'm thinking about has a comparator and then some way of turning the result into a logic circuit to decide what to turn on (or off).

Thoughts?

How would one design a circuit to do that? Each light bulb would presumably have it's own relay.

Look at post #12 (provided by person #12)
1) did it with current transformers
2) That is one of the inputs to the comparators in that circuit (the comparators are the triangles).
3) ditto
4) Designing the load array is easy enough. I did the drawing with triacs instead of relays.
5) Light bulbs are by far the cheapest way to buy a load resistor.

Thoughts? Look at the drawing.

 

4. Based on this difference send signals to relays to turn on, or off, a sequence of light bulbs (if you used 6 halogens, 300W x 4, 200W x 1, 100W x 1, you could add additional load to the line in 100W increments and you could get every value from 100W to 1500W with the fewest # of bulbs, without exceeding 300W bulbs, to keep cost and heat down).

Keep in mind you have a useful coincidence here where 1+2+4+8=15 so a minimum set of binary weighted series of bulbs meets your max power needs. It's great you know of 200 and 300W bulbs (I was assuming 100W bulbs assuming the gov't keeps from enforcing the law banning them) but higher watt bulbs may be easier to use.

I think the minimum set of bulbs would be:
100W: 1 bulb
200W: 1 bulb
400W: 2 bulbs
800W: 3 bulbs

for 4 switches and 7 bulbs total

5. A small fan to blow over the bulbs while it is running, just to cool things a bit.

Unit the whole room hits equilibrium <grin> A 3,000 watt load can get very hot. I'd say put it outside but that could be a lot of light pollution on pretty summer evenings.

How would one design a circuit to do that? Each light bulb would presumably have it's own relay.

Yep. Relay, triac, some sort of switch per bulb bank, meaning when you turn the 400W load on one switch turns on 2 bulbs.

 

Yanmar YDG-5500. Picked it up a few weeks after Huricane Sandy. As I am not really using it, I should be testing it monthly and running it on load for at least a half hour. While there is much discussion on load banking in giant gen sets, there are no design reasons I can think of why the same principles should not apply to my 5K. I need the 5k because despite my average use of about 1.8KW (wintertime; no AC), I have two sumps and must be able to accomodate the inrush current on their 8A motors.

Do you use the 240 V output?