Snow melt system design

[Kecarbaugh]

I’m in process of designing a snowmelt system for my outbuilding apron and entry door landing. Given the price of tankless water heaters being cheaper than a boiler, I thought that I could possibly utilize a modulating circulation pump, changing the flow, to externally modulate the tankless water heater since the model that I selected (Rinnai RU199IN) internally modulates based upon inlet and set point temperature and flow rate. I am also proposing that I only have a single primary pump and not a primary/secondary setup as this is a dedicated heater. Is this a reasonable and effective solution? Am I missing some important design considerations?

NOTE: see revised design later in this thread...

 

[WorthFlorida]

I know nothing about your project other than it can get quite expensive on fuel cost, however, if it prevents someone from a slip and fall it is worth it.

One item is check with Rinnai that the heater box and elements are OK with an anti freeze solution. A Indirect water heater maybe needed as a heat exchanger if Rinnai doesn't support it (at least for warranty purposes).

Check with Rinnai if the tankless needs a minimum water pressure before it will fire up. You may need to add a back flow prevent and a pressure reducing valve to keep constant pressure. The attached picture is from my brother's VT home. A propane tankless that heats the home and feeds an iIndirect water heater for domestic hot water. VT water is about 45ºF out of the well.

 

[Sponsor]

 

[Kecarbaugh]

I know nothing about your project other than it can get quite expensive on fuel cost, however, if it prevents someone from a slip and fall it is worth it.

One item is check with Rinnai that the heater box and elements are OK with an anti freeze solution. A Indirect water heater maybe needed as a heat exchanger if Rinnai doesn't support it (at least for warranty purposes).

Check with Rinnai if the tankless needs a minimum water pressure before it will fire up. You may need to add a back flow prevent and a pressure reducing valve to keep constant pressure. The attached picture is from my brother's VT home. A propane tankless that heats the home and feeds an iIndirect water heater for domestic hot water. VT water is about 45ºF out of the well.

Thank you for your reply. Rinnai does allow a glycol mixture without invalidating the warranty. However, as part of the self modulation capability the flow will be restricted so that the outlet temp stays at the selected setting of 140 degrees, which I believe will inhibit sufficient flow through the primary loop to enable sufficient snowmelt. Looks like I'll have to go with a boiler to get the proper results.

 

[Kecarbaugh]

So I've revised my proposed design so that I use a boiler instead of a tankless water heater in order to get better modulation control. I'm looking at the Rheem RCBH199DVLN combi boiler which will also have the advantage of providing DHW (which I originally planned a separate tankless unit for that), and a boiler heat loop for the snowmelt system. A couple of concerns about this setup. I'm concerned that the 120k BTU output of the boiler heat loop may not be enough heat due to my elevation of 8500 feet and 40% glycol reducing the heat efficiency of this system. I calculate that the heat load is around 82k BTUs which would provide 200 BTUs/sq ft. Snow free ratio of 0.5. I'm also concerned that the built-in circulator within the combi boiler may not be sufficient for the amount of head ~19ft to induce sufficient flow. Revised design diagram attached. Any thoughts?

 

[wwhitney]

However, as part of the self modulation capability the flow will be restricted so that the outlet temp stays at the selected setting of 140 degrees, which I believe will inhibit sufficient flow through the primary loop to enable sufficient snowmelt.

I'm newish to hydronics, but I'm not following this statement. Once you've designed your snow-melt emitter loops (which I don't know how to do), for a given outdoor condition and fraction glycol, the emitter is going to have a given heat output rate as a function of the incoming water temperature and the flow rate through the loops.

For a fixed incoming water temperature to the emitter, the effect of flow rate on heat delivered is non-linear, and you reach a point of diminishing returns. This article explains it nicely: https://www.pmengineer.com/articles/93344-flow-rate-heat-output-and-delta-t

For changing incoming water temperature, I'm not 100% sure of the behavior, but my first thought is that the heat delivery rate will be close to proportional to the temperature differential between incoming water temperature and the load temperature.

Anyway, my understanding is that if you can design your emitter to provide the necessary heat output with 140F incoming water and a reasonable flow rate, and you can provide a pump that will give you that flow rate given the head losses in the system, then as long as the resulting heat delivery rate is below the tankless's rating (possibly adjusted for the glycol fraction, if that changes how well the tankless's internal heat exchanger extracts heat from the flame), it should be able to give you 140F water at the required flow rate.

Cheers, Wayne

 

[John Gayewski]

Given your loop lengths are so drastically different a delta p pump is I think what your after. That will keep the flow rate the same on loops with different head pressures. That is the purpose of a delta p pump.

 

[Kecarbaugh]

I'm newish to hydronics, but I'm not following this statement. Once you've designed your snow-melt emitter loops (which I don't know how to do), for a given outdoor condition and fraction glycol, the emitter is going to have a given heat output rate as a function of the incoming water temperature and the flow rate through the loops.

For a fixed incoming water temperature to the emitter, the effect of flow rate on heat delivered is non-linear, and you reach a point of diminishing returns. This article explains it nicely: https://www.pmengineer.com/articles/93344-flow-rate-heat-output-and-delta-t

For changing incoming water temperature, I'm not 100% sure of the behavior, but my first thought is that the heat delivery rate will be close to proportional to the temperature differential between incoming water temperature and the load temperature.

Anyway, my understanding is that if you can design your emitter to provide the necessary heat output with 140F incoming water and a reasonable flow rate, and you can provide a pump that will give you that flow rate given the head losses in the system, then as long as the resulting heat delivery rate is below the tankless's rating (possibly adjusted for the glycol fraction, if that changes how well the tankless's internal heat exchanger extracts heat from the flame), it should be able to give you 140F water at the required flow rate.

Cheers, Wayne

My point was that the glycol mix running through a tankless water heater initially has the delta T of potentially 148 degrees if near the freezing point of the glycol mix. The tankless water heater would trickle the water through until the cement slab warmed up, and as your referenced article points out the heat transfer drops off significantly at low flow rates. I think it would be better if the flow rate was not modulated by the tankless water heater and it just does the best it can at a higher flow rate, even though it doesn't reach 140 degrees on the output. With my new design the modulation is performed by the snow melt controller, and the flow rate would be constant. I'm just concerned that the circulator is not powerful enough to push the glycol mix with ~18 ft of head. Trying to find out more about the circulator from Rheem.

 

[wwhitney]

My point was that the glycol mix running through a tankless water heater initially has the delta T of potentially 148 degrees if near the freezing point of the glycol mix. The tankless water heater would trickle the water through until the cement slab warmed up, and as your referenced article points out the heat transfer drops off significantly at low flow rates.

Ah, I was only thinking of the steady state conditions, while you are concerned with the start up conditions.

But I still don't see the problem. The Rinnai unit will do 375 degree F * gpm for pure water. 40% glycol has a specific heat of 85% to 90% that of water (edit: not sure if that's by mass or by volume, so if it's by mass, the density change should go in here), depending on temperature, so call that 420 degree F * gpm for your fluid. So what's the worst case ground temperature you'd see at a cold startup, 0F? If the Rinnai is set to 140F outlet water, then it would throttle the flow rate to 3.0 gpm under those conditions.

That's below your target of 7 gpm for your emitter. So your emitter is going to be emitting less heat during startup. But with only 9 gallons of fluid in the system, all the fluid will have circulated through the tankless after 3 minutes at 3 gpm. Now the incoming fluid temperature (to the Rinnai) isn't 0F, but whatever the emitter return temperature is at 3 gpm flow with 140F. So the Rinnai will increase the flow it allows as the return temp increases. Once the return temp hits 80F it should do 7 gpm. And then if your pump is going to limit you to 7 gpm, if the return temp climbs above 80F, the Rinnai will start modulating its burner.

So startup doesn't seem like a problem to me. But again, I'm new at this.

Cheers, Wayne

 

[Kecarbaugh]

Given your loop lengths are so drastically different a delta p pump is I think what your after. That will keep the flow rate the same on loops with different head pressures. That is the purpose of a delta p pump.

I do have separate balancing and isolation valves at the manifold which will partially help address the issue, although with the significant difference with the one zone, I'm not sure what impacts this will have. I will be turning down the flow rate significantly in the short zone and will take some tweaking to make sure the system distributes the heat evenly.

 

[Kecarbaugh]

Ah, I was only thinking of the steady state conditions, while you are concerned with the start up conditions.

But I still don't see the problem. The Rinnai unit will do 375 degree F * gpm for pure water. 40% glycol has a specific heat of 85% to 90% that of water (edit: not sure if that's by mass or by volume, so if it's by mass, the density change should go in here), depending on temperature, so call that 420 degree F * gpm for your fluid. So what's the worst case ground temperature you'd see at a cold startup, 0F? If the Rinnai is set to 140F outlet water, then it would throttle the flow rate to 3.0 gpm under those conditions.

That's below your target of 7 gpm for your emitter. So your emitter is going to be emitting less heat during startup. But with only 9 gallons of fluid in the system, all the fluid will have circulated through the tankless after 3 minutes at 3 gpm. Now the incoming fluid temperature (to the Rinnai) isn't 0F, but whatever the emitter return temperature is at 3 gpm flow with 140F. So the Rinnai will increase the flow it allows as the return temp increases. Once the return temp hits 80F it should do 7 gpm. And then if your pump is going to limit you to 7 gpm, if the return temp climbs above 80F, the Rinnai will start modulating its burner.

So startup doesn't seem like a problem to me. But again, I'm new at this.

Cheers, Wayne

Wayne,

I am new to this too. I see your point about transitioning to steady state will eventually allow the desired flow rate. If I went the route of letting the Rinnai unit modulate the flow based upon it's delta T, then the snow melt controller would try to modulate the flow as well and I'm not sure how well this would work. The snow melt controller is designed to not thermally shock the concrete upon startup, but I suspect the modulating flow from the tankless unit would help perform the same function to an extent. Seems the flow restriction in the tankless water heater would increase head pressure at least initially. The flow charts point out the dependence on altitude, delta T and set temperature and says the flow chart is only valid for incoming water temps is < 70 degrees, but my attempts to get more information from Rinnai for temps above 70 resulted in the response of "I need a boiler for my application as the unit cannot support a incoming temperature > 100 degrees". Not sure if the inlet temp would ever exceed 100 degrees as I would think the snow melt controller would intervene and start restricting the flow as the cement slab heats up, effectively lowering the inlet temperature.

With the boiler design, the modulation is with the temperature, not the flow (although I believe there would be still a relationship there as I need more information from Rheem on what the behavior is when it can't achieve the desired temp).

Ken

 

[wwhitney]

"I need a boiler for my application as the unit cannot support a incoming temperature > 100 degrees".

So how did you do the emitter design, and does it really need 140F water to function? Seems like with a snow melt system you'd never want to raise the heated area above, say, 35 or 40 F, and without knowing anything about how snow melt emitters are designed, I'm surprised you would need water 100F hotter than the target temp.

Anyway, I just mention this in case there's an advantage to being able to use a Rinnai instead of a boiler. In which case if you can set the Rinnai to, say, 110F and get your emitter to work with a return temp that is always 100F or less, the above objection would not apply.

Cheers, Wayne

 

[Kecarbaugh]

So how did you do the emitter design, and does it really need 140F water to function? Seems like with a snow melt system you'd never want to raise the heated area above, say, 35 or 40 F, and without knowing anything about how snow melt emitters are designed, I'm surprised you would need water 100F hotter than the target temp.

Anyway, I just mention this in case there's an advantage to being able to use a Rinnai instead of a boiler. In which case if you can set the Rinnai to, say, 110F and get your emitter to work with a return temp that is always 100F or less, the above objection would not apply.

Cheers, Wayne

I used an online tool from a website. It actually specified 146 degrees, but the Rinnai unit can only be set to 120 or 140 degrees. The design also was based upon a 25 degree drop during steady state, so setting it to 120 would fit within the parameters of the Rinnai unit, but might not deliver as many BTUs per the design The Rheem boiler actually only allows 120k BTUs for the snow melt system, which might be a little low considering that I'll loose 20% or so at my altitude and maybe another 10-15% due to the glycol. The advantage with the Rinnai unit is that it connect via Wi-Fi with an optional accessory, while the Rheem boiler cannot. The Rheem boiler has the advantage that I would only need one device for both DHW and snow melt and would save money, otherwise I was planning on two dedicated Rinnai units.

Ken

 

[John Gayewski]

This all sounds hokey to me. I've never heard of modulating the flow based on temp and I don't see the purpose.

Flames modulate for a reason. Circulators modulate for different reasons.

 

[Kecarbaugh]

So the attached diagram represents what I've come up with so far. Due to the amount of head, I've added a circulator that has a reasonable flow rate at the ~19 feet of head I'm expecting. It won't get to the 7 gal/min unless I use two of the specified Taco high head pumps in parallel. Is that advisable, or should I just get a bigger single pump?. Due to the circulator within the combi boiler I've implemented a primary/secondary configuration to obtain hydraulic separation. I've also included check valves and purge valves to the diagram. Open to any suggestions or things I've missed.

Ken

 

[John Gayewski]

So the attached diagram represents what I've come up with so far. Due to the amount of head, I've added a circulator that has a reasonable flow rate at the ~19 feet of head I'm expecting. It won't get to the 7 gal/min, but should be a reasonable 4 gal/min. Due to the circulator within the combi boiler I've implemented a primary/secondary configuration to obtain hydraulic separation. I've also included check valves and purge valves to the diagram. Open to any suggestions or things I've missed.

Ken

Are you not using a fill valve for the hydronic side?

 

[Kecarbaugh]

Are you not using a fill valve for the hydronic side?

Thank you, I will add that.

Ken

 

[WorthFlorida]

Interesting on your approach and planning.
1) What is your expected temperature to the (assuming) concrete to melt ice and snow?
2) Does the melted run off have a place to go before it re-freezes?
3) Is there a recommended maximum temperature to a frozen, say 20ºF, concrete slab so its does not crack?

 

[Kecarbaugh]

Interesting on your approach and planning.
1) What is your expected temperature to the (assuming) concrete to melt ice and snow?
2) Does the melted run off have a place to go before it re-freezes?
3) Is there a recommended maximum temperature to a frozen, say 20ºF, concrete slab so its does not crack?

The slab temp is controlled by the Tekmar 654 controller, but I suspect it would provide enough heat to raise the temp of the concerte slab to be around 35 degrees or so to adequately melt the snow. The boiler can be set up to 180 degrees by the controller. The melted snow would run off the concrete apron onto a gravel driveway made of decomposed granite which drains pretty well. Not sure what is the criteria that would keep the slab from being 'shocked'. Again, the Tekmar 654 will control the temperature and is suppose to ramp up slowly to prevent damage to the slab.

Ken