Hacker News Reader

Comment by sokoloff | Hacker News Reader

Sort of. My house is 100 years old and was built without insulation (some rooms were insulated through remodels over the years). It uses water circulation and cast iron radiators and stays comfortable with my ancient cast iron boiler and 180°F water.

To know whether a heat pump (air-to-water) can replace that boiler effectively and maintain comfort, I had to find out whether the house would be comfortable with water temps of 135°F or so. Is there an amount of “that’s just looking at the building construction” to make that analysis? I think maybe technically yes, but practically no.

As it was, to get an answer, I abused my old boiler by turning the water temps down (causing condensation and slow damage [planning to replace it anyway]) and seeing what happened on cold days.

View on HN

Don’t you want to replace the radiators with large vertical floor to ceiling radiators to get the same amount of heat transfer at a lowest possible temperature?

I am not familiar with the nuances, but this appears to be an equation with two variables - water temperature and radiator surface area. Maximising surface area should slow you to use lower water temperature. And the lower the water temperature, the more efficient the heat pump will be?

As a mechanical engineer, would I like to do that? Absolutely!

As a homeowner who bought a 1920s house because I like the character of a 1920s house, would I like to do that? Absolutely not!

You didnt need to do that, though. It peobably wouldve been easier to use manual j (or some software) to estimate the heat loads in the house, with given set points, using weather station data.

I can find/closely estimate the heat loss easily. What is much harder to find is the heat gain/transfer into the room from 1920s cast iron radiators at 135°F flow and the balance of the system flow temperatures at those 45°F lower flow temps than originally designed.

Then, because the answer is almost always going to be "yeah, it's going to be really close...", I felt well-advised to prove it via experimentation rather than commit to changing the heating plant to a system that could not provide 150°F flow temperature.

If you are ok with IP units the formula is 500 * GPM * delta-T for water as the fluid.

The question is not about the general hydronic heating formula [nor manual J heat loss estimations], but rather "what will the delta T of the rad in this particular room, in this piping network [it's a converted gravity feed system, now being a pumped], using 69°F room temp and 135°F leaving water temp from the heat source?"

This is going to be a complex problem because of the shape of the radiator and you'll need to calculate radiative and convective components of heat transfer i.e. you'll need finite element analysis to do this. If you simplify it to a simple shape like a rod or slab you can get somewhere in a calculation, but this is only going to give an instantaneous measure because of heat transfer to the rest of the universe.

Alternately, to get a realistic measure, you'll need to set your boundary conditions about what the heat flow out of the room will be, which is a bit simpler to setup with U-values, area, delta-T, and heat capacity of materials. You'll also need to do this to every other room in the building simultaneously. This is a Manual J, or heat balance method or the radiant time series method load calculation that will balance out with the amount of heat leaving your radiator without knowing its specific shape.

Right. Or, I can turn a screw down on the aquastat on a cold day and see what happens, which is what I did (a handful of different times).

For the, for lack of a better word, standard radiators there is a formula with a dT^4. But I totally agree, this isnt all that straighforward, given that for the dozons of installers, experts and home owners I have spoken, Ive heard dozens+1 methods for estimating. Estimating heat loss from a given building and estimating power output of a given installation of radiators, very few people seem to be able to calculate that.

If your heater can go low (mine bottoms out at 50 unfortunately), by far the easiest is to just test.

What was the answer you came to? I'm in a very similar situation

Ref: https://news.ycombinator.com/item?id=39144329

A two-part answer:

From the Mechanical Engineering/Thermodynamic angle:

> Ultimately, I proved to myself that a heat pump could work down to an outside air temp of about 18°F [which is slightly above our 99th percentile design temp] with flow temps of 135°F, so an air source heat pump could work with slightly reduced comfort on about 2% of days or could work all the time with supplementation with a 9kW [30K BTU/hr] electric boiler.

From the commercial angle:

> What killed the project is no heat pump installer was interested in doing the work (as reflected by outright declining to bid, while bidding a 4-hour gas boiler swap, or by bidding so high that they might as well not have bid, while also cheerfully bidding a 4-hour gas boiler swap). So my house still burns gas for heat.

On the engineering front: I think the answer is often going to "<hissing inhale> It's going to be close; we should probably test it..."

Thanks for the response. My feeling is that maybe in about 10 years Air-to-water Heat pumps will be more common in the US and we might have a better chance of getting a reasonable installation quote

Reply on news.ycombinator.com