@jamespa Here's an update to my previous post. Mitsubishi engaged a subcontractor to investigate the poor performance of my heat pump. He concluded that it was defective and that it should be replaced. The replacement was installed by another Mitsubishi subcontractor on 15 December. There is a very significant improvement in performance. However, it is still nowhere near meeting its specification. Initial indications are a COP of 3 at 9C OAT with a flow temperature of 35C and a flow rate of 19l/min. Mitsubishi quote a COP of 5.1 at 7C OAT with a flow temperature of 35C. Still a long way off meeting their specification.

Posted by: @ecoste

↑

There is a very significant improvement in performance. However, it is still nowhere near meeting its specification. Initial indications are a COP of 3 at 9C OAT with a flow temperature of 35C and a flow rate of 19l/min. Mitsubishi quote a COP of 5.1 at 7C OAT with a flow temperature of 35C. Still a long way off meeting their specification.

Improvement is good.  Lets explore where we are now.   You should definitely be getting better than 3 at 9/35  (Can you confirm 35 measured at and by the heat pump, ditto COP, or are they independent measurements if so where and how?).  That said mine at least seems to take a while to settle down.      Is it cycling?  If so with what on and off times.

Posted by: @ecoste

↑

@jamespa Here's an update to my previous post. Mitsubishi engaged a subcontractor to investigate the poor performance of my heat pump. He concluded that it was defective and that it should be replaced. The replacement was installed by another Mitsubishi subcontractor on 15 December. There is a very significant improvement in performance. However, it is still nowhere near meeting its specification. Initial indications are a COP of 3 at 9C OAT with a flow temperature of 35C and a flow rate of 19l/min. Mitsubishi quote a COP of 5.1 at 7C OAT with a flow temperature of 35C. Still a long way off meeting their specification.

I've shared this table in one of my posts recently, as I have the same unit as you, but the 5.1 COP you refer to is only one number quoted within their table, and is the highest figure quoted within these, under the "Partload2" element of the testing, when the capacity is restricted to 5.8kW.  The COP figures range from 3.67 to 5.10 at those temperatures within the table and these are from tests carried out in laboratory conditions, so about as 'real-life' as the MPG figures quoted for vehicle fuel efficiency.

This isn't to say that there's no room for improvement, as I've been investigating the COP of my own system and am in the process of trying various means to tweak any further COP efficiencies from it.  The closest 'real-life' reference point I can provide at present is from 15th December, where my system had a reported heating COP of 3.42 on a day with average OAT of 9.2°C and an average flow rate of 33.88°C when running under weather compensation that day (figures provided via the Mel Pump app).  These also fall short of any of the reported figures in the table, but these are being run on a retrofit installation, with 'real-life' performance inhibitors having an influence on what the true maximum is.  I suspect there may be further tweaks available that will have a positive impact, but my gut feel for now is that 3.8 is likely a realistic maximum for my installation at those temperatures.

I would be inclined to disregard the 5.1 COP figure as being anything close to realistic.  I found the table to be really helpful as a means to re-base my expectations in terms of questioning "what good looks like" for my own specific circumstances.

Good to hear that the faulty unit has been replaced.

@jamespa No cycling, flow temp varies by less than 2 degrees, return is very stable. Today it is running at 38C varying by less 2C, since water heating finished at about 5.00 this morning. So about 12 hours continuous running with OAT of 8C. It is providing about 3.5kW of heat from 1.3kW of electrical energy, COP of 2.7. Temperature measurements are close to the cylinder in the loft, about 20m from the heat pump. They are taken from the Mitsubishi controller, but are integer numbers, i.e. no decimal places so not great granularity. the flow rate is also taken from the Mitsubishi controller, still integer values. Electricity is taken from the Powervault App, allowing 300W for house usage, this is a generous figure based on observations over the last 2 years.

As an addendum to what @jamespa said, can you also let us know what the deltaT is as well, and whether or not there’s any defrosting going on (unlikely at that kind of OAT, but worth asking the question).

Oh, and @ecoste, I don’t suppose by any stretch that you’re running Home Assistant, are you?

Posted by: @ecoste

↑

@jamespa No cycling, flow temp varies by less than 2 degrees, return is very stable. Today it is running at 38C varying by less 2C, since water heating finished at about 5.00 this morning. So about 12 hours continuous running with OAT of 8C. It is providing about 3.5kW of heat from 1.3kW of electrical energy, COP of 2.7. Temperature measurements are close to the cylinder in the loft, about 20m from the heat pump. They are taken from the Mitsubishi controller, but are integer numbers, i.e. no decimal places so not great granularity. the flow rate is also taken from the Mitsubishi controller, still integer values. Electricity is taken from the Powervault App, allowing 300W for house usage, this is a generous figure based on observations over the last 2 years.

OK 12 hours stable running should be enough to get a good reading.

Can you be more precise about where the flow and return temperature sensors which are being used to calculate COP are?  Is there a buffer tank/LLH between them and the heat pump?  What are the actual figures, because 1C error (integers only) can make quite a difference if its running at a DT of (eg) 2!  Also temperature sensors are notoriously inaccurate, have you swapped them round/are they well attached to the pipes and insulated on the outside and/or provided with thermally conductive paste.   Is it glycol or water in the system,

If its delivering 3.5kW at 19l/min means a DT of 2.7 (assuming pure water).  Small errors in temp measurements can make a big difference to calculated COP/power.

Also according to the datasheet whats the expected COP under min load (3.5kW is near to min load)

Posted by: @ecoste

↑

@jamespa No cycling, flow temp varies by less than 2 degrees, return is very stable. Today it is running at 38C varying by less 2C, since water heating finished at about 5.00 this morning. So about 12 hours continuous running with OAT of 8C. It is providing about 3.5kW of heat from 1.3kW of electrical energy, COP of 2.7. Temperature measurements are close to the cylinder in the loft, about 20m from the heat pump. They are taken from the Mitsubishi controller, but are integer numbers, i.e. no decimal places so not great granularity. the flow rate is also taken from the Mitsubishi controller, still integer values. Electricity is taken from the Powervault App, allowing 300W for house usage, this is a generous figure based on observations over the last 2 years.

My observations on this are less than 24 hours old, but the system flow rate reporting did something unexpected yesterday as part of this experimentation.

The system has consistently reported a delta T of 2-2.5°C, which a bit of background reading suggested was a bit lower than ideal.  In an attempt to see what would happen I reduced the pump speed setting, which was set at the maximum of 5, to run at a pump speed of 3, to see if that had any impact.  The delta T subsequent to this does look to have increased to 3.5-4°C when the heating is running, which is the output that I had expected.  However, the flow rate reporting hasn't noticeably changed from the pump speed being reduced, and continues to show 24-25 l/min.  Logically, this doesn't make sense in the context of the delta T having increased, so I'm currently unsure how accurate the Ecodan flow reporting data is.  Regardless of what the system is doing it's consistently showing either 24-25l/min when running or zero when idle, with nothing much between the two at any time.

It's brand new data to me, so I can't conclude too much from it, but I'd expect to have seen both numbers change, not just one of them.