How to control primary flow rate of plate exchanger based on DHW demand

[sirocosm]

Yes, I would put the thermal balancing valve on the return of the primary (store) side. I think the pump could go on either side.

I don't think you need anything to sense the temperature on the output of the DHW, it is assumed to be too hot anyway and will be mixed down with a TMV, so it also doesn't matter if it varies a bit.

The problem I see with a sensor on the DHW side of the exchanger is that you can't really sense its temperature unless there is flow through it. When there is no hot water demand, the sensor would go cold, open up the primary, and it would run for no reason. You could use some type of special exchanger with a sensor in it, but I would prefer off the shelf parts that are easy to replace.

 

[hard-work]

The pump is always put on the cool side of a plate heat exchanger.

You now throw in a TMV on the DHW, which you never mentioned, which will control the DHW setpoint. These mechanical valves do not react too quickly. The pump may "hunt", giving an erratic delivery of hot water into the plate. .

Having:

1. The smart pump on the primary side after the plate.
2. In-line 2-port thermovalve, right on the outlet of the plate.
3. A DHW TMV on the DHW outlet.

This "may" work, but I doubt it. However, you still have two control mechanisms out of sync with each other - the smart pump and in-line thermovalve which will most likely cause pump hunting. Then a third control mechanism in the TMV thrown in.
What temperature would the thermovalve on the primary side be set to?

 

[echoes]

It stands a better chance of behaving, and stability, if the pump ran at a constant speed, but you would need a 'shunt' for when the thermostatic valve closed, and a flow switch.

Then you have one sensor (the thermostatic valve's remote sensor on the DHW out, before any TMV) controlling one transducer (the thermostatic valve on the primary).

Any latency in feedback (and there will be some) will reduce stability, so minimise this by design.

 

[sirocosm]

The idea is to leave the pump on all the time, and get it to reduce its output when the valve closes. Having the pump on all the time reduces the amount of control elements required (less to go wrong) and modulating the pump should reduce the flow so it doesn’t destroy the stratification (and waste power) when it is not needed.

It works really well on my central heating circuit. I have a smartpump (Alpha 2 set to modulate on pressure) and TRVs on all radiators, and no wall stat. As the TRVs open the pump automatically increases its speed. All my rads have the lockshields wide open, and there is no need to “balance” the system in the traditional sense, the TRVs do the balancing. This system is very nice as the radiators are never “off”, unlike most systems where you have no heat at all when the boiler is off. The only downside is you cannot use a programmable main thermostat because there isn't one, you need programmable TRVs all around if you want to automatically turn the heat down at night.

I have never seen a case of oscillation in my setup, even when only one rad is turned on.

 

[hard-work]

If say the thermal store is set to 80C, what will you set the 2-port thermovalve to?

• The best location for this 2-port valve would be in the inlet pipe from the thermal store cylinder, directly above the plate.
• Cylinder setpoint is say 80C.
• 2-port thermovalve set to say 75C.

Cylinder up to temperature. The smart pump pumps water through the thermovalve, it reaches 75C then closes.
The temperature will drop, so the thermovalve opens slightly to maintain the plate temperature at around 75C.

DHW is drawn off, with cold water running through the fresh water side of the plate. This will extract the heat from the primary side of the plate cooling it right down.
The thermovalve opens allowing hot, 80C, primary water through the primary side of the plate. (this will take time).
The operation would most likely be, immediate hot water at the taps, then a slug of cold water because the 2-port thermovalve is slow to open and react. Then there will be hot primary water running through the plate at 80C which will close the 2-port thermovalve. So then only cold water. It will not work.

Put the 2-port thermovalve directly on the outlet of the plate's primary side, then the temperature will be less than its setpoint of say 75C, as the heat is transferred into the cold incoming water, keeping the valve open. Better chance of working.
Worth a try if you want to go that route, with much experimenting in the setting up of the setpoints of the store of water and 2-port valve. I would install an oversized plate to ensure reduced, or no, slug of cold water and to ensure the old mains water takes all or most of the heat in the plate.

 

[sirocosm]

I would set the store a bit lower than that, say 60C, which is what I run mine at. Most people run hotter because their exchangers are too small, mine is 4.1 square meters surface area, so 60C works fine. I am guessing one could set the thermal valve around 40C. When the tap is opened cold water will drop the return temperature, since the directions of flow are reversed between the primary and DHW sides. i.e. the feed side of the DHW is the return side of the primary, and the feed water will always be lower than 40C.

"Put the 2-port thermovalve directly on the outlet of the plate's primary side ... " This is what I am suggesting, to put it on the return of the primary.

I don't think the valves are as slow as you think, otherwise they would not work for showers. There will always be a gradient along the length of the exchanger i.e. from one end (cold mains in/primary return) to the other (hot water out/primary feed). Initially the water may indeed be hotter, because the cold will be warm for a bit, but the TMV after that should take care of it.

One thing that might be an issue is that you might need a different target temperature for the termovalve in summer vs winter, when the input water is cooler.

 

[hard-work]

You can buy fast or slow acting 2-port valves. These are usually from expensive makers like Danfoss generally aimed at the commercial market. You have to look at their data sheets. A fast acting 2-port valve causes the smart pump to react, then the 2-port valve closes as more heat from the heat store is pumped in, then the smart pump again reacts. This may cause hunting (up and down) - a big fear. A slow acting 2-port valve may cause cold water slugs.

The only way is to try, and experiment with setpoints, and see what happens. It may work. The fact you have a large plate makes matters better. Because makers have not used this approach on their products does not mean it will not work with some tinkering.

 

[hard-work]

The problem I see with a sensor on the DHW side of the exchanger is that you can't really sense its temperature unless there is flow through it. When there is no hot water demand, the sensor would go cold, open up the primary, and it would run for no reason. You could use some type of special exchanger with a sensor in it, but I would prefer off the shelf parts that are easy to replace.

You can fit a tee on the outlet of the DHW on plate heat exchanger. A temperature probe pocket can be in the tee as far as you can get it into the plate heat exchanger. A 2-port thermo valve with a remote sensor then goes in the pocket sensing the DHW more accurately inside the plate. The DHW side of the plate will be hot as heat is transferred from the primary side. A 15mm by-pass around the pump can be fitted with an in-line isolator valve. This can be just cracked open to allow gravity circulation through the plate from the thermal store to keep it warm.

 

[sirocosm]

That sounds like it should work as well, although I don't think you would need the bypass. If the valve closes that is fine. As soon as the remote sensor drops in temperature, the valve would open up and the pump would increase speed.

The sensor would be on the output of the DHW side. It could then also replace the TMV function as well.

One down side would be that if the tank temperature dropped below the target DHW temp, then the pump would run until the store heated up, even if the hot water was off. This would not be a problem with the sensor on the primary.

 

[hard-work]

Simple pipe/cylinder stats can hold off pumps. But more complexity with more setpoints to tinker with to get the setup right.