Weather Compensation and Defrosting

[ReedRichards]

Today it's a little below zero outside, and foggy.  My heat pump is having to take frequent breaks from heating the house to defrost itself.  But I have never seen a heat pump a with means of taking this into account when operating in Weather Compensation mode; do these exist?  If the heat pump was metering its own heat input into the house then some sort of Defrost Compensation should be possible.  This would also allow more accurate compensation for time off heating the DHW cylinder, which could vary from day to day.  

[PhilT]

Interesting question - a quick search suggests maybe not. I run my HP in target temp (auto adaptive) mode which I guess is a kind of workaround - if DHW and/or defrost cause a fall in room temp, the flow temp automatically compensates.

[JohnMo]

I believe Nibe take this into account. I think they alot a quantity of heat and then make up for defrosts. Could be wrong.

[HughF]

Yet again another reason why these things need load compensation…

 

I’m so tempted to start on that opentherm - modbus converter I was thinking of building/coding up.

[Beelbeebub]

1 hour ago, HughF said:

Yet again another reason why these things need load compensation…

 

I’m so tempted to start on that opentherm - modbus converter I was thinking of building/coding up.

What is the physical layer for opentherm? It seems to be some sort of 2 wire bus.  I haven't seen any opentherm modules on the hobby market like you see for CAN or rs485

[JohnMo]

Just compensate with flow temperature

 

Two givens when using a heat pump without a sizable buffer.

 

When the heat pump cannot modulated any further it shutdown and when it defrosts, it shutdown.  Both events cause the flow temp within the heating system to drop. Therefore the mean flow temperature is lower over the heating period.

 

So just add a degree or so to the heating curve to fix it or add a buffer.

 

Nothing complex needed, 30 seconds to move the heating curve.

[ReedRichards]

13 hours ago, JohnMo said:

Nothing complex needed, 30 seconds to move the heating curve.

Yes, sure.  I have a Weather Compensation curve that is supposed to cover the full range of temperatures but I actually have to manually intervene to move the curve when the temperatures get low enough that the heat pump requires frequent defrosting.  I could even set up an outdoor thermometer that sets off an alarm whenever the temperature outside gets too low. 

 

Either something is properly automated or it isn't; I don't want to be part of the automisation.  

[JohnMo]

9 minutes ago, ReedRichards said:

but I actually have to manually intervene to move the curve when the temperatures get low enough that the heat pump requires frequent defrosting

Isn't defrost a given, when you get down to about 4 deg? So why keep moving settings, once you have made a change isn't it just left where it is for the next time?

[HughF]

21 hours ago, Beelbeebub said:

What is the physical layer for opentherm? It seems to be some sort of 2 wire bus.  I haven't seen any opentherm modules on the hobby market like you see for CAN or rs485

Not sure what the physical layer is, I think it’s something custom, but it’s designed to re-use old thermostat wiring.

 

quite a few master and slave shields/interfaces are available. I might grab one. The plan would be to read the target flow temperature, calculated by the load compensating opentherm thermostat, from the thermostat. Convert this into a modbus flow temperature packet for the target heat pump, and then send it to the heat pump.

 

A web front end (it would connect to your home wifi) would allow for customisation/selection of the heatpump from the list of those supported.

 

 

[joth]

2 hours ago, HughF said:

Not sure what the physical layer is, I think it’s something custom, but it’s designed to re-use old thermostat wiring.

 

quite a few master and slave shields/interfaces are available. I might grab one. The plan would be to read the target flow temperature, calculated by the load compensating opentherm thermostat, from the thermostat. Convert this into a modbus flow temperature packet for the target heat pump, and then send it to the heat pump.

 

A web front end (it would connect to your home wifi) would allow for customisation/selection of the heatpump from the list of those supported.

 

 

So gain load compensation at the loss of weather compensation? Or do OpenTherm stats support both?

 

Homely is the other option . be interested to here some real world reviews of it

https://midsummerwholesale.co.uk/buy/homely/homely-smart-controller-V3

(No ecodan support yet, else I'd already have got one)

 

[Beelbeebub]

2 hours ago, HughF said:

Not sure what the physical layer is, I think it’s something custom, but it’s designed to re-use old thermostat wiring.

 

quite a few master and slave shields/interfaces are available. I might grab one. The plan would be to read the target flow temperature, calculated by the load compensating opentherm thermostat, from the thermostat. Convert this into a modbus flow temperature packet for the target heat pump, and then send it to the heat pump.

 

A web front end (it would connect to your home wifi) would allow for customisation/selection of the heatpump from the list of those supported.

 

 

I'm fiddling with my UFH (boiler driven) - at the moment it is just on/off from the thermal store at a set temp.

 

I'm planning on measuring external temp (via my MHRV intake), and doing a simple linear weather comp to get the desired delta T between the floor slab temp and the desired internal temp.

 

Its then easy to calculate the desired slab temp (thermostat setting plus calculated delta T). 

 

The controller then activates the UFH pump until the slab return temp is to the target.  When it is, it shuts a valve that controls the hot water input into the mixer, but keeps the pump circulating to measure the slab temp.

 

Rinse repeat 

 

 

[jayc89]

My gas boiler doesn't support load comp so I'm just running weather comp. At the minute the flow temp is peaking around 60c, so slightly above the condensation point, UFH mixer runs at a max of 40c, but we need the higher temp for the upstairs radiators, at least until I finish insulating those rooms. UFH has been running pretty much all day this week, radiators upstairs are on for a couple of hours a day and we have priority hot water where the flow temp goes to max when the cylinder calls for heat (which is approx 1 hour a day)

 

I'm thinking about wiring the upstair rad stat the same way as the cylinder stat, so when either the hot water or rads call for heat, the flow temp is max, would mean I could reduce the weather comp curve to better suit the UFH. I don't think we'd run above the condensation point for that much longer either as typically the cylinder calls for heat whilst the rads are on anyway. So perhaps 2 hours running at a hotter temp, but likely 4-6 hours, during this current weather, running at a much lower temp when only the UFH is calling for heat. 

[JohnMo]

1 hour ago, jayc89 said:

At the minute the flow temp is peaking around 60c, so slightly above the condensation point,

Just so you are using the correct information - Condensation point is measured on the return water temperature ,not the flow water temp.

[JohnMo]

2 hours ago, Beelbeebub said:

The controller then activates the UFH pump until the slab return temp is to the target.  When it is, it shuts a valve that controls the hot water input into the mixer, but keeps the pump circulating to measure the slab temp

Why not do away with the conventional UFH mixer and just use a 3 port mixer with actuator to give you full control of the flow temperature to the floor. A normal UFH mixer always mixes a percentage of return water into the hot water. A 3 port mixer doesn't. Reuse the pump and top adapter piece to the manifold.

 

Instead of running the pump 24/7 could a simple room thermostat be used to give a feedback loop? Or run a pump sniffer cycle for 10 mins an hour.

[HughF]

4 hours ago, joth said:

So gain load compensation at the loss of weather compensation? Or do OpenTherm stats support both?

 

Homely is the other option . be interested to here some real world reviews of it

https://midsummerwholesale.co.uk/buy/homely/homely-smart-controller-V3

(No ecodan support yet, else I'd already have got one)

 

If you’re going to have a thermostat (and yeah, most people will want one), it needs to set the target flow temperature rather than being a crude on/off stat. So you might as well revert to load compensation.

 

Turn it up, rads get warmer, turn it down, rads get cooler….

[Beelbeebub]

3 hours ago, JohnMo said:

Why not do away with the conventional UFH mixer and just use a 3 port mixer with actuator to give you full control of the flow temperature to the floor. A normal UFH mixer always mixes a percentage of return water into the hot water. A 3 port mixer doesn't. Reuse the pump and top adapter piece to the manifold.

 

Instead of running the pump 24/7 could a simple room thermostat be used to give a feedback loop? Or run a pump sniffer cycle for 10 mins an hour.

I might, but it requires a load more plumbing.

 

What I actually have is a standard thermal actuator (a UFH zone valve) plumbed into the return from the 4nport mixing valve for the UFH.

 

This is historical to preven some thermosyphon from my thermal store to the UFH.

 

Currently it is plumbed in parallel with the UFH pump. When the thermostat calls for heat, the pump is activated and the zone valve (eventually) opens.

 

Until that happens the UFH water just recirculates round the slab.  As the valve opens the mixing valve starts to blend and the flow temp rises to whatever it's set to.

 

I've noticed that I can actually control the flow temp to any point between slab return temp and the mixing valve setting by opening the zone valve fractionally.

 

As the valve moves slowly, my idea is to measure the flow temp and then "blip" the thermal actuator on and off to hold it at the intermediate opening point that gives the desired flow temperature.

[Beelbeebub]

The pump will only run when the room temp (measured by a probe in the MHRV room extract manifold) is below the set point.

 

Eventually some logic will be put in so that continually running the pump whilst the slab return is at target temp (which is taken to indicate the target slab temp is too low to maintain the desired temp) will trigger a slight uplift in the target slab temp.

 

(Maybe - if I can be arsed to do all the work)

[Beelbeebub]

The big thing slowing me up is that the current system is very simple, uses no custom hardware and actually works pretty well.

 

Here is the internal temp track for the last few days.

 

Realistically, i'm chasing marginal gains.

 

(But it is a fun exercise)

 

 

[HughF]

Mine is just ‘on’…. The pump is wired onto a 13a socket.

[Beelbeebub]

47 minutes ago, HughF said:

Mine is just ‘on’…. The pump is wired onto a 13a socket.

The pump runs 24/7?