@majordennisbloodnok  forecast.solar is still dead for me. I've uninstalled and reinstalled, and all the values still appear as 0.

Posted by: @morgan

↑

@majordennisbloodnok  forecast.solar is still dead for me. I've uninstalled and reinstalled, and all the values still appear as 0.

I use "Solcast PV Forecast"

Posted by: @morgan

↑

@majordennisbloodnok  forecast.solar is still dead for me. I've uninstalled and reinstalled, and all the values still appear as 0.

Sorry, @morgan; I forgot to reply to this. Is it still dead for you?

If yes, are you able to (independently of HA) get to their website at https://forecast.solar?

@majordennisbloodnok yes, still dead. Shame becauase it was a handy way to tell if/when I would need to set a schedule to charge my batteries from the grid. And via HA was very convenient.

Posted by: @morgan

↑

@majordennisbloodnok yes, still dead. Shame becauase it was a handy way to tell if/when I would need to set a schedule to charge my batteries from the grid. And via HA was very convenient.

Hmm. I agree; a pain.

And the link to their website? Does that work for you from your home?

If the website access works, it may just be that the API key you generated may have expired and need regenerating. Worth checking the website access first, though.

@majordennisbloodnok it isn't obvious where on their site one finds where to get a forecast.

Posted by: @morgan

↑

@majordennisbloodnok it isn't obvious where on their site one finds where to get a forecast.

That’s not a problem. If you can hit the main landing page (which you obviously seem to have managed) that’s all I need to know. It confirms there’s no problem between your home and the web server.

Having read up more, the API shouldn’t need a key at all. I don’t know if it used to but doesn’t now or what, but I’ll need to get back home to see what I can see from my home network before going any further.

I thought this might be the most appropriate place to reply to the post by @sheriff-fatman here

For the last couple of weeks I've had access to extensive system data via the MelPump app and, in particular, the data available via the dongle supplied by @f1p which I've also managed to feed into Home Assistant (the rapid learning curve associated with using HA for the first time has been a 'fun' project to get to grips with) and I'm rapidly improving my understanding of how the heat pump is functioning at any given point in time.  In particular, I've gained an appreciation of how difficult it is, in reality, to get any meaningful output data from a heat pump, beyond those based on calculations using flow rate and delta T.  Short of spending £700 or so to retrofit monitoring equipment on the flows, I now have access to real-time data reported by the heat pump to do these calculations, but it has been eye opening to realise that there's no way of actually having any metered measurement of this, and to realise that any reported COP figure from any system is a function of whatever calculation that app is doing behind the scenes, so they're all educated guesses of one form or another.

The adapter isn't perfect in estimating the energy figures, it does depend on the resolution and calibration of the Ecodan's onboard sensors for output power and input power estimates based on the standard compressor sizes and relationship with flow temperatures

I pitch the £35 adapter as giving users the data and visibility to make the "big gains", where there are design or commissioning issues preventing the system from operating as well as it could.

The several hundred pound monitoring kits (plus the potential requirement to modify pipework) then can provide the accuracy of 0.1C on temperatures and 0.1l/min flow rates to give those fine tuning "minor gains".

Posted by: @f1p

↑

I thought this might be the most appropriate place to reply to the post by @sheriff-fatman here

For the last couple of weeks I've had access to extensive system data via the MelPump app and, in particular, the data available via the dongle supplied by @f1p which I've also managed to feed into Home Assistant (the rapid learning curve associated with using HA for the first time has been a 'fun' project to get to grips with) and I'm rapidly improving my understanding of how the heat pump is functioning at any given point in time.  In particular, I've gained an appreciation of how difficult it is, in reality, to get any meaningful output data from a heat pump, beyond those based on calculations using flow rate and delta T.  Short of spending £700 or so to retrofit monitoring equipment on the flows, I now have access to real-time data reported by the heat pump to do these calculations, but it has been eye opening to realise that there's no way of actually having any metered measurement of this, and to realise that any reported COP figure from any system is a function of whatever calculation that app is doing behind the scenes, so they're all educated guesses of one form or another.

 

The adapter isn't perfect in estimating the energy figures, it does depend on the resolution and calibration of the Ecodan's onboard sensors for output power and input power estimates based on the standard compressor sizes and relationship with flow temperatures

I pitch the £35 adapter as giving users the data and visibility to make the "big gains", where there are design or commissioning issues preventing the system from operating as well as it could.

The several hundred pound monitoring kits (plus the potential requirement to modify pipework) then can provide the accuracy of 0.1C on temperatures and 0.1l/min flow rates to give those fine tuning "minor gains".

 

Just for complete clarity on this, the comment quoted wasn't intended to imply any criticism of the dongle, which I think is a brilliant piece of kit.  The comment is written from the gradual realisation over time that the expectation as an 'uneducated' user of a heat pump that it will be able to accurately measure energy inputs and outputs from the system isn't actually true, yet so much of the discussion around heat pumps focuses on CoP scores, which by default require accurate measurement of both.

The reality, which I only really fully grasped from configuring the Home Assistant setup, is that while there's a pretty good chance that electrical input can be recorded accurately, and can be monitored via a digital meter or a Shelly-type CT clamp, the output side of things can only ever be a heat-based calculation as there's no electrical output generated that can be metered in the same way.  I would assume that there is also a margin of error (to what extent I have no clue) from the flow rate and Delta T calculations, even if you're using one of the expensive monitoring kits to get numbers that are as accurate as possible to feed into the calculations.  With the dongle, I have the next best thing to this for my own system, in terms of accurate data, but at the end of the day whichever method of measurement is being used is still only providing numbers to be fed into a theoretical formula.

Given this, I find it somewhat odd that the narrative on heat pumps is so heavily driven towards reporting CoP and SCoP, as the end user is essentially at the mercy of whatever formula is being used by a particular manufacturer or app, the details of which are generally hidden behind the scenes.  In fact, the cynic in me thinks that there's a tangible benefit for heat pump manufacturers to ensure that their systems over-report these scores for marketing purposes.  The scenario that immediately springs to mind is seeing something from an Aira system user, who was seeing longer term CoP scores of over 7 being reported, and questioning the accuracy of this in his commentary given that the best scores he'd seen for any system on Open Energy Monitor was closer to 5.  He sensibly concluded that he either had the greatest performing heat pump system in the world or, as seemed more likely, there was something fishy within the reported numbers.

@majordennisbloodnok thank you. A great explanation. But why is this so complicated? How do we expect mass adoption when customers need a degree in data science?

Posted by: @carpenterstation

↑

@majordennisbloodnok thank you. A great explanation. But why is this so complicated? How do we expect mass adoption when customers need a degree in data science?

Totally agree.

The problem lies with what I was once told was the WIIFM factor - What’s In It For Me?

Mitsi create a heat pump and get an installer to set it up for you. There’s a value to them to get your data so they make sure the installer knows how to set up a wifi dongle. There’s no value to them to let you control the heat pump locally so they don’t bother adding extra connections to their kit to make it easy.

Attaching an extra bit of third party kit to your heat pump is a way to make control other than via Mitsi technically possible, and using Home Assistant to talk with those bits of third party kit is often the least bad way of converting “technically possible” into real life practical operation. However, a sensible plug and play solution requires Mitsi to adopt industry standards there’s no WIIFM factor for them to do so.

I should point out that wherever I’ve mentioned Mitsubishi it’s perfectly valid to substitute in any other manufacturer’s name; this is an industry wide issue.

@sheriff-fatman — a good assessment of the relative pros and cons of what I call 'good enough' vs ultra rigorous monitoring. I have gone with the former, the cost and invasiveness of the letter did not seem worth it to me, but that is not for a moment to say others shouldn't use it if they want to do so.

I do my monitoring using python scripts with modbus over a wired connection. The data collection is pretty bombproof, but the data itself all comes from the manufacturer's sensors, in my case Midea, and there is ample opportunity even if not taken for them (and any other manufacturer) to 'mark their own homework'. That said, I do have independent checks for some key variables, including energy in (independent kWh meter that supplies only the heat pump) and flow rate (in line analogue meter fitted at installation time). All in all, taking these independent data sources into account, I think I can say my monitoring is 'good enough'.

I think the focus on COP comes from the fact it is a single performance number, like, though not the same as, miles per gallon. We naturally want our kit to perform well, and COP provides a summary measure of that. It can also be used as a first indicator that something is wrong eh a COP of 2 when it should be say 3-4 is an indicator something is wrong, which of course be in the monitoring, or the heat pump or both. But I have also said that the most important number is not the COP, it is the IAT, because that is a proxy for human comfort, which cannot yet be monitored over modbus. No doubt someone somewhere is working on an implantable cortical electrode with external A and a B terminals, but I won't be fitting one anytime soon. As long as my IAT is in the right range, and my COP is tolerable, I can live with that. 

Edited to correct typos