Grundfos Alpha 2L Proportional Pressure

[D_Hailsham]

The problem is that, with a fixed speed pump, when the flow reduces, e.g because a rad is shut off by its TRV, the head increases. But the resistance of the actual circuit reduces because the flow has reduced. If you recalculate the index circuit you will find that the head has reduced.

I have a spreadsheet which models the performance of my system. I have found, by experiment, that if the output of any rad is set to zero, which effectively is the same as shutting it off, the resistance of the index circuit reduces. So the flow rate reduces and the required pump head reduces.

Resistance increased because the rads closed, so you only need enough water for one radiator

No! The resistance of the circuit reduces but the pump head increases because the flow has reduced.

 

[picasso]

if your doing multiple zones have multiple pumps, you can tailor every zone to its requirements.

 

[fezster]

One thing that bothers me about the primary circuit of the LLH is the use of a fixed speed pump. All the documentation says to use a pump which overcomes the maximum resistance of the boiler (in my case, 4.05m at 38KW).

In reality, though, the boiler spends most of it's life modulated down, and thus the flow requirement is usually much lower. I'm unsure what will happen in reality though - the flow rates stated in the manual are MINIMUM flow rates, so my assumption is the boiler will still operate correctly, but perhaps be less efficient (delta T at the boiler will be lower perhaps?).

EDIT - Thinking about this in more detail, this would only be true if there was no circulation in the secondary circuit. With heat being transferred across the LLH, the boiler should maintain a delta T of 20 by modulating up and down dependent on heat required by the secondary circuit.

 

[John D v2.0]

Resistance increased because the rads closed, so you only need enough water for one radiator

No! The resistance of the circuit reduces but the pump head increases because the flow has reduced.

I'm confused. If you have 10 rads you have 10 parallel paths through the system for water to go. Close 9 of them, and all the water squeezes through one rad. Therefore overall resistance of the system increases, and (with a fixed speed pump) both the flow and head across the remaining radiator increases.

With a PP pump adjusted for that situation, the head across the overall system would reduce by control of the pump, so that the head across the remaining radiator would stay exactly the same. But that probably isn't an exact science.

 

[D_Hailsham]

David was kind enough to do an index calculation for me a few years ago, for each of my zones, but the reality is somewhat different as I've discovered by using flow and return temps to determine the actual flow rates in my system (and therefore the actual resistance of the system too).

Apart from knowing the flow and return temps you also need the boiler output at that time. How did you obtain this information?

How far out were my calculations, and were they over or under the actual figures?

 

[fezster]

Apart from knowing the flow and return temps you also need the boiler output at that time. How did you obtain this information?

How far out were my calculations, and were they over or under the actual figures?

I measured gas consumption over a period of time (I think it was 5 mins). Not the most scientific way, I know, but I had no way of knowing the actual output the boiler had modulated to. (EDIT - Just to add, I did this same procedure multiple times, using differing d0 outputs at the boiler. I was able to confirm gas consumption increased and decreased accordingly, and was able to see Delta T increasing and decreasing accordingly too, always arriving at the same flow rate - this was to prove to myself that my method was correct).

I redid your calculations (went more conservative) and the figures came to 1.6M upstairs and 2.4M downstairs. My downstairs circuit has a second pump (long story, but it is what it is for now), so difficult to work out what the actual pump curve is. So I measured upstairs in isolation, and this is what I came up with:

34kw (used gas consumption to confirm)
21 delta T (taken from d40 and d41)

Flow rate = 0.39 l/s (1,404 l/h)

Grundfos 25-80 speed 3 - 6.85m head

View media item 100913
ecotec plus 438 head loss - 3m head

View media item 100912
Actual head loss in system (upstairs only) - 3.85m

 

[fezster]

if your doing multiple zones have multiple pumps, you can tailor every zone to its requirements.

I have considered this. The way my pipes are laid out, it lends itself well to putting a LLH between the primary flow/return and the Zone Valves. If I was to have each zone piped separately from the LLH, there'd be considerable more pipework alteration involved. I also don't like the idea of having to run multiple pumps if I can avoid it. I do get your point, though.

 

[fezster]

I'm confused. If you have 10 rads you have 10 parallel paths through the system for water to go. Close 9 of them, and all the water squeezes through one rad. Therefore overall resistance of the system increases, and (with a fixed speed pump) both the flow and head across the remaining radiator increases.

The flow rate for 10 parallel paths would be the sum of each of the individual path's resistances. If you reduce those to one single path, then *at the same flow rate* the resistance of that ONE path would be much higher than it was when it was 1 of 10 paths (more water being pushed through that single pipe). Of course, you don't actually need that flow anymore (assuming you want to maintain the same delta T across your radiator), so the actual resistance of that path is less if you calculate it at the required flow rate. I think this is what David means by the index circuit reduces as rads are closed.

Now if we talk in terms of what's actually happening in the real world - the pump is pushing water through 10 paths at a given flow rate. Reduce that to one path, and the pump now tries to push the same flow rate through one path (because it's a fixed speed pump). The resistance is much higher than it was when pushing through 10 paths, so the pump follows it's curve upwards - i.e. the flow rate reduces, and the pump differential pressure rises.

If the pump was smarter, it would realise that the flow doesn't need to be so high for that one path, so it can reduce it's output and push a lower flow rate at a lower differential pressure instead. Hence, a variable pump is able to reduce it's head and achieve the same required flow rate.

All of that makes sense in my head. It could be complete nonsense though!

 

[John D v2.0]

the resistance of that ONE path

Maybe he did mean that, but that whole chain leads back to this post you made which was talking about the pump that covers the whole system:

I'm sure most of that is wrong (hence the question marks above), because if resistance increases, surely more head is required?

However although I understand the theory, I feel like people are mixing/misusing terms, so there's a good chance we're all violently in agreement but under different terminology!

 

[D_Hailsham]

We appear to have a true "chicken and egg" situation here! In the case of a fixed speed pump, when a radiator shuts down, does the flow reduce because the pressure increases, or does the pressure increase because the flow reduces? The pump can only react to the conditions, it doesn't dictate what happens.

As I said earlier, if a radiator is shut off the system's flow requirement reduces, and so does the pressure loss of the index circuit. A fixed speed pump cannot meet this requirement, but a variable speed one can. Whether it can do it precisely, i.e provide the exact correct flow and pressure, is another matter.

 

[fezster]

or does the pressure increase because the flow reduces?

So the pump head increases because the flow reduces - as the pump is operating on it's fixed curve?

The pump can only react to the conditions, it doesn't dictate what happens.

This is what I was referring to when I said

Now if we talk in terms of what's actually happening in the real world - the pump is pushing water through 10 paths at a given flow rate. Reduce that to one path, and the pump now tries to push the same flow rate through one path

A fixed speed pump knows nothing about the flow rate requirements in the circuit. It pushes X number of litres/s at P pressure differential. If we assume you are correct (and I've no doubt you are) and the index circuit decreases, this is at the REQUIRED flow rate, which is less than X. But at X, the resistance is now higher (it has to be as you have one path of water, whereas previously you had more than one). And as you said "the resistance of the circuit varies proportionally to flow squared", so the flow must reduce (following the pump curve). Obviously in real terms, all of this happens concurrently resulting in the flow finding a happy medium to the resulting resistance. And that happy medium is a reduction in flow, but at a greater pump head.

A "smart" pump using PP alleviates this by intentionally lowering it's flow rate to try to match the system resistance.

EDIT - Again, I'm not saying any of the above is actually correct. I'm actually asking the question if my understanding of how it works is correct?

EDIT2 - I changed some of the above because Im sure what I wrote originally was wrong.

 

[John D v2.0]

In the case of a fixed speed pump, when a radiator shuts down, does the flow reduce because the pressure increases, or does the pressure increase because the flow reduces?

When a rad closes the resistance increases, that's all.
Due to the pumps characteristics, both the flow reduces and the pressure increases, but that's down to the pump not the rad. Flow/pressure doesn't cause the other.
It's analogous to a power supply, could be constant voltage or constant current, but most supplies are somewhere between the two. So take a partial load off the power supply and the voltage will go up and the current down. But voltage and current didn't cause each other to change.

if a radiator is shut off the system's flow requirement reduces, and so does the pressure loss of the index circuit.

The radiator shutting off is just the simplest way of a radiator forcibly signalling to the system that it doesn't have any heat requirement any more. This is pretty brutal, and leaves the pump and boiler relying on side effects of the act to decide what to do.

The side effects with a fixed unregulated pump and unregulated boiler are increased head, reduced flow, and warmer flow temperatures from the hex.
The pump may and the boiler must then modulate themselves to keep things stable.

 

[John D v2.0]

A pump using PP alleviates this by intentionally lowering it's flow rate to try to match the system resistance.

Yes, and as a thought experiment a far more accurate but over kill and expensive way would be to place a pressure sensor at the single branching point for all rads in both flow and return (ie manifold). All rads would have to branch at the same point.
Then the pump would modulate it's speed purely on the pressure difference on those sensors. Then the flow rate would always be the same through all the rads.
Anything else is an approximation, but probably a fine one for practical purposes.

 

[fezster]

Anything else is an approximation, but probably a fine one for practical purposes.

I think a lot of what is going on is probably solvable by complex equations, but it'd be nice to have a broad understanding of what is going on within the system with regards to pressure and flow rate.

 

[fezster]

https://www.diynot.com/diy/threads/...portional-pressure.495124/page-2#post-4063202

I measured gas consumption over a period of time (I think it was 5 mins). Not the most scientific way, I know, but I had no way of knowing the actual output the boiler had modulated to. (EDIT - Just to add, I did this same procedure multiple times, using differing d0 outputs at the boiler. I was able to confirm gas consumption increased and decreased accordingly, and was able to see Delta T increasing and decreasing accordingly too, always arriving at the same flow rate - this was to prove to myself that my method was correct).

I redid your calculations (went more conservative) and the figures came to 1.6M upstairs and 2.4M downstairs. My downstairs circuit has a second pump (long story, but it is what it is for now), so difficult to work out what the actual pump curve is. So I measured upstairs in isolation, and this is what I came up with:

@D_Hailsham You never came back to me on this post. It'd be nice to get validation on whether my method of empirically working out the system resistance was correct?