Multiple boilers in P/S, sizing their circulators

Luc

Brad, thanks for all the good advise. It is really helpful for a novice in hydronics like me.

I looked at the Wilo Stratos pump curves and they all require a minmimum of about 3' of head. Anyway, electricity here is at a good price and I think the payback would be quite long.

I will give a quick look at the Grundfos variable speed.

For the moment I am leaning towards the Grundfos Superbrute UPS43-44FC for the boiler circulator pump. With a balancing valve, I will be able to get the right flow.

Thanks,
Luc

Luc

We are doing some changes in our hydronic heating system. The present 7 cast iron oil boilers are all piped in parallel. We are getting rid of 4 boilers and replacing them by an off-peak electrical accumulator and an electrical boiler. There will be some controls to keep our peak electrical demand constant.

We want to pipe all the boilers as secondary into the primary loop. There will be a total of 5 boilers. Three of the original cast iron oil boilers and the two electrical boilers. The cast iron boilers will be on standby for emergency use or during high demands.

I have a few questions on selecting the circulators for the boilers.

1)What are the best boiler circulators? A circulator with a flat pump curve or a steep pump curve?

2) I want to use 1.5" or 2" pipes and circulate at about 30GPM. Is it realistic to circulate at around 30GPM in 1.5" black steel pipes? They usually reccommend 22.9GPM for copper.

Thanks,
Luc

Brad White_200

As Luc would have it

I have some thoughts for you.

1. In proper P/S piping, the effect of the secondary (what I call the radiation/distribution circuit) on the primary or boiler circuit, should be nil. Negligible. Thus, once set up, a given circulator running at a certain flow and head point, should not change in any meaningful way regardless of what the secondary circuit is doing.

Steep or flat? Depends on the pressure drop of the boiler it serves. If a high pressure drop, the curve will be steeper and go to a higher head. Low pressure drop boilers such as cast iron conventional types are really minimal. If you use one foot, that is probably too much. Pump curves serving such boilers would be rather flat.

The notion of "Flat versus Steep" is really germane to systems which have a variable dynamic such as control valves or variable resistance. With P/S, your boiler circulators should have a constant resistance regardless of what it's neighbors are doing.

Don't forget though,your pump selection in any case should be made to your flow and pressure drop.

2. For the pipe size, I personally would go to 2-inch because I like to keep my pressure drops low for parasitic losses. You have to operate the boiler pump by default so why not minimize that if you can? Then again, it is not my money I am spending here.

That said, you should still crunch the numbers in due diligence and mindful of the budget if nothing else.

At that flow and assuming that you have 50 total equivalent feet of piping, fittings, valves, etc. in the near-boiler piping, your head loss with 2-inch would be less than one foot, call it 0.9 feet actually. The velocity would be 2.86 FPS.

If you have a foot for a cast iron boiler, that is barely 2.0 feet of head.

May be hard to find a circulator for that low a head, unless you go with a Wilo or Grundfos variable speed circulator. More money but you get exactly the pump you want, despite anyone's ability to calculate it.

If you use 1.5-inch pipe, the head rises to 3.13 feet and the velocity would be 4.73 FPS. Add a foot for a cast iron boiler and you have call it 4.2 feet on a good day. You will likely find better selections given that you have at least some head to work against.

Depending on your actual fabrication, I can go either way. A short run even at a high velocity does not scare me, unless the job is noise-sensitive.

Luc

Brad, thanks for the answer.

I did a rough calculation this morning for the circulator on the electric boiler and it came down to about 2.75' of head loss at around 23GPM for 1.5" pipe. I was really surprised that it was this low. Whitout crunching the numbers for the cast iron boilers loops, I thought that they would probably have an even smaller head loss. You confirmed it.

Now, what does everyone do when they have to include multiple boilers? On paper it looks good to hook up multiple boilers as a Secondary but if we have a hard time to find a circulator with low head and high flow what do we do?

Luc

Brad White_200

Circulators

Is your flow 23 GPM or 30? I imagine it varies by boiler and the required temperature rise. In the end, it is all temperature rise anyway.

If your budget allows.... well, let me put that aside for a second. If it does not work, your budget does not matter. Let's go with what will work.

Try a Wilo Stratos. I think a 1.25 x 3-20 (the smallest), might suit you but Mark Hunt might confirm that. The variable speed feature gives you nearly infinite curves to balance your flow and head requirements. Head is mostly a function of RPM. Once you can generate the flow at a high head, lower the RPM to get the real head you need (very simplistically stated, I submit).

The "old fashioned way" is to install the smallest circulator which will meet the flow requirements and which, in your case, would have a surplus of head. Once installed, you would install a balancing valve to choke it down to the actual flow and head requirements. Driving with the brakes on.

You can still do that but it will cost more to operate. Crunch the numbers though. The difference in net HP used versus the higher number of a constant speed circulator has to be brought into dollars and cents (CDN in your case!) in order for an owner to make a decision.

Once each boiler is so tuned (manually for a constant flow at head), the "distribution loop" to the radiation/emitters, has a circulator sized for that duty. If that duty includes control valves, a dynamically responsive variable speed circulator is what I would use. As the load changes so does the pumping capacity and electrical usage goes down by the cube root of flow. Such a deal.