If I have read and copied the Midea data tables correctly, I am not sure I can see much 'efficiency' difference between a smaller heat pump running flat out and a large heat pump running at a lower output. The data here is from Midea's 8kW foot warmer (on the left) vs their 16kW roaster (on the right). Copying PDF tables is always a hit and miss affair, but these tables are just about readable, and what you want to do is, comparing across the same OAT and LWT, find matching outputs (the CAP or capacity column (in Watts), the other two are COP, and PI which is power input in Watts), and then look at the COP and PI values. Of course, none of the outputs match, making this very much a Mark I Eyeball exercise, but in general the 8kW COPs are perhaps slightly worse than the 16kW for similar outputs. For example, at an OAT of 10 and a LWT of 40, the 8kW unit can put out 7774W at a COP of 4.48 using 1735W (the 'norm' row) while the 16kW unit can put out 7441W at a COP of 4.59 using 1620W (the 'min' row) (relevant cells highlighted in attached image). The outputs are not identical, with the 16kW model output slightly lower, and the difference in COP is small, and furthermore these are 'lab' results, ie controlled conditions, not typical real world conditions.  

I would much rather have observational data from real world installation...   

A few Mitsubishi figures from their data book for comparison... 

5kW output at 12C OAT:

5kW unit is COP of 6.15

11.2kW unit is COP of 5.9

5kW output at -7C OAT:

5kW unit is COP of 3.0

11.2kW unit is COP of 2.9

The 11.2kW is the closest to double the 5kW in Mitsubishi's R32 range, the 6kW is a software limited 8.5kW so not a good comparison.

Comparing the Mitsubishi 8.5kW to the 14kW for an output of 8kW, at 12C and -7C, gives the same COP values.

So a larger unit running at approx. 50% vs smaller unit at approx. 100% doesn't seem to always result in greater efficiency. 

I feel the best first question after establishing that a heat pump is over-sized (even massively so say 3x) is does that affect efficiency?

John Cantor covers this well in this YouTube session https://m.youtube.com/watch?v=0oszaqP0TI4  

the critical aspect being is it at the limit of its turn down range and then short cycling (which is inefficient). If it isn’t inefficient then the being over-sized is fine, but often isn’t.

Then forcing the cycling to effectively a shorter time on period would be a good first tactic. This would work best with high thermal mass, eg. UFH or solid brick house with good insulation. Aligning the on time to cheap electric is an opportunity.

Posted by: @judith

↑

Then forcing the cycling to effectively a shorter time on period would be a good first tactic. This would work best with high thermal mass, eg. UFH or solid brick house with good insulation. Aligning the on time to cheap electric is an opportunity.

Pretty much I'm coming to that conclusion, TBH it was my hunch when I started this thread but I wanted (and still want) to see if anyone comes up with anything different.  So far not but it's still open to debate.

Posted by: @jamespa

↑

Posted by: @judith

↑

Then forcing the cycling to effectively a shorter time on period would be a good first tactic. This would work best with high thermal mass, eg. UFH or solid brick house with good insulation. Aligning the on time to cheap electric is an opportunity.

Pretty much I'm coming to that conclusion, TBH it was my hunch when I started this thread but I wanted (and still want) to see if anyone comes up with anything different.  So far not but it's still open to debate.

I agreed with your approach on the other thread and so pulling out the general case is a better way of capturing it.

On LG ashp they have a "silent mode" that operates at reduced power (intended to reduce noise at night) and some users use successfully to reduce cycling etc in oversized pumps. 

I dont know if other ashp have similar function?

@tim441 Re the silent mode, It made me think have a missed something, but i have looked it up, on the Viessmann all the silent mode does is restrict the fan speed, this doesn't reduce the minimum capacity of the compressor therefore doesn't effect minimum output. When the pump is trying to go as low as it can, to reduce output, the fan is already at a reduced speed.

@mikef interesting! 

As mentioned .. from info in LG user group it does reduce cycling and output. But I don't use it myself as my heat pump seems approx right size.

On this forum with advanced monitoring available a Mitsubishi user whose pump was oversized recommended the cloud based auto-adaptation function to improve COP.

It tends to more technical discussions there