Solar Reporting

[Horsley1]

Hi
Very basic question, but how can I get/download the data for the amount of solar power my panels are producing? I can see the real time data from the box but in the app and on the site, I am unsure which number it is. All I want is the daily total of production as it will help me understand how much electricity I am using on a daily basis. I get my grid total from Octopus but am missing the balancing number.
Thanks

PS - I have solar and 4 Foxess batteries

[Luke]

On the cloud site there is a "Total Yield" under Sites > MySites which will tell you total gen since your site has been communicating with FoxESS Cloud service. Doubt it'll be 100% accurate with data missing due to the cloud service being out or your own internet going offline.

As for seeing historical data, you can goto Device > Inverter > More Options Icon. It has a default dasboard for the stuff you're looking for then at the bottom you can select something like "PV1Power" and it'll show you a graph for whatever day you have selected. I've not seen a way to export this data though

Personally I've setup HomeAssistant with the InflixDB and Grafana addon and query the data myself via modbus over ethernet. From here I don't need to rely on the FoxESS cloud service to be online and hold my data and grafana is very flexible in how you can model dahsboards and reports

[Horsley1]

Luke
Thanks for the reply. Following your guide, I can see the totals you describe in the overview section and have taken a monthly view and then just highlighted the Total Yield bars which I can then read. So, not downloadable but gives me an idea of what my panels have produced. and I will just look at it every now and again and add it to my spreadsheet
Ray

[Horsley1]

Will

Thought I had updated this with more info but doesn't seem to show - so apologies if it is duplicated).

I have a 4 battery solution (the V2400 I think) so that adds up to 9.6kWh. I charge these up every night at the cheap rate. ie, it should be for about 8.7kWh since they start at 11%. I notice that, when I am awake at 5.30, the total yield number is zero. However, as the batteries drain down (and before the solar kicks in at 9.30am), the total yield number rises as the battery drains so that, when it is empty (well 11%) the total yield number is 7.8kWh. Given other comments about the true fill-ability of the batteries, this seems to make sense as 7.8 is about 90% of 8,6. Then during the day, the total yield rises very slowly, it is winter after all, and suggests that I am getting around 3-6kWh at the moment as the total yield rises to around the 11kWh mark during the day as the panels kick in (I have 5.8 kWh of panels).

Is that a reasonable assumption? Do others find this as well?

Regards

Ray

[SolarJunkie]

Hi Ray,
I’ve been looking at this recently, after finding my total yield figures seemed out since using a cheap overnight rate to charge the batteries.
I can see no way to verify the solar yield, as the charge and total yields include, but do not differentiate between solar and grid charge. Also, the total yield seems to include discharge in the equation too.
I have contacted Fox in the last two days and am awaiting a response. It is possible to interrogate the inverter directly, but it only seems to give a figure for the day and a grand total for all time.
It is, to my mind, quite an important omission.

[Horsley1]

Solar Junkie

Any response back from FoxeEss? I'm still dubious that my maths is anywhere near accurate

Ray

[Dave Foster]

If you are using the Fox App you can work out the solar generated from the 'Monthly Yield' statistics pages, the formula is -

Solar generated = (battery charge - battery discharge - consumption yield + Load yield + Feed in yield)

note the Load yield is found on the Site Overview Monthly Yield graph, the rest are on the Device Monthly Yield graphs

[SolarJunkie]

If you are using the Fox App you can work out the solar generated from the 'Monthly Yield' statistics pages, the formula is -

Solar generated = (battery charge - battery discharge - consumption yield + Load yield + Feed in yield)

note the Load yield is found on the Site Overview Monthly Yield graph, the rest are on the Device Monthly Yield graphs

Ah! The missing piece of the puzzle, I couldn't, for the life of me, find what I needed, ie the load.

Thank you once again.

Is there information anywhere that explains what each of the graph parameters mean, not the individual sensor metrics, just the sum totals of the overviews?

For example, the difference between Load and Consumption

[Dave Foster]

Ah! The missing piece of the puzzle, I couldn't, for the life of me, find what I needed, ie the load.

Thank you once again.

Is there information anywhere that explains what each of the graph parameters mean, not the individual sensor metrics, just the sum totals of the overviews?

For example, the difference between Load and Consumption

There’s nothing written down in the user manuals, but it’s probably been asked a few time in the Facebook group - hopefully we’ll be able to build up the resources on here for others to find.

• Consumption yield is everything you have imported from the grid which will include battery charge, house load, EV charger etc…

• Load consumption is only the house load (which include solar, battery and the grid)

• Total yield is everything that has come from the hybrid inverter (either solar or battery)

• FeedIn yield is what you export to the grid

[Horsley1]

Dave

This is so useful, I really appreciate it. However, the FoxEss app that I have (naturally) has different words. I have tried to match with your words against what my version of the app says:

Feed-in-yield - same - so what goes direct from the solar to the grid (not via the battery). This is nil (and has been all winter !!)

Total Yield - same - so everything that goes through the inverter. Can I check that this will be 3 things:
- solar that goes straight from the panels to the house
- solar that goes from the panels to the battery (because the solar is producing more than the house needs at that moment in time). This goes through the inverter
- everything that goes from the battery to the house. This will include both the power that I have downloaded from the grid (at my cheap rate) as well as any topups that have been done by the solar panels as per bullet above. So this means that there would be some double counting?

Consumption Yield - same - so all grid imports - battery charge/house/EV, etc. Interestingly, this figure never matches what Octopus charge me eg for 3 random days - 82.7 vs 85.12, 47.3 vs 48.9 and 60.5 vs 62.27. The Total Yield number on the monthly stats page is always the same as this so is it the same thing?

Battery Charge - is this the same as Charge Energy Total that I have (so charge from grid as well as solar panels)

Battery Discharge - is this the same as Discharge Energy Total that I have? And, given that my battery always discharges fully every day at the moment, why isn't this the same as Battery Charge? There always seems to be a 0.7 to 1.3 difference with the charge energy total being higher

Given what you've said in your calculation (repeated below) and what I have above, then currently, my solar is simply the difference between the 2 battery numbers because

Solar = 10.7- 9.5 - 50.9 + 50.9 + 0 = 1.2

Solar generated = (battery charge - battery discharge - consumption yield + Load yield + Feed in yield)

If this is correct, I am disappointed that my panels (5.77kWh) are only producing between 1 and 5 kWh - even with the poor weather and incessant rain

Hope you understand (and sorry for the questions - but it is bloody confusing.

Ray

[asyrop]

Hi Ray

Firstly, I want to thank you for posting the orginal query and flushing out the various answers along the way - it is really useful for me and I'm sure other newbies who have much the same queries.

Secondly, as I try to make sense of my data, I thought a few additional comments might help. In fact I am trying to find out why my H1 inverter / Battery Controller do what they do as it seems to vary with time and maybe temperature (my first post in Dec). This is not the thread to raise these questions, but I have been struck by my installer who reckons to have done 60+ of these Fox systems and I am the first to raise questions as to why the behaviour is both strange and beyond my control. It seems Fox has a lot of explaining to do including to their direct customers installing their kit !

So in response to your last post:
a) I don't think the Fox reported numbers are that accurate. I have tried frequently to get them to correspond on an hour by hour or even daily basis. I personally reckon their firmware brings together many pieces of data and generates rounding errors or assumes they are steady between sampling when in fact they will have been varying. Will has long pointed out that the main app does this with snapshots of PV power values at 10 mins intervals when they system should integrate the power over the period to give the average power or the reported energy produced with much greater accuracy. I can well imagine it does the same in many other parts of the firmware and/or app software. If anyone knows different, I would love to hear !!

b) My installer included a generation meter which gives imported energy / exported energy and the difference to the Solar PV devices from my consumer unit. I thought it unnessary at the time but now see the value. Over a period I can see my net solar gain from the panels (FYI 70kWh since 20th Nov from 5.2kW peak panel on low pitched roofs). On a daily basis I have to allow for the change in SoC of the battery between readings, but at least I have a reliable measure (which I have to record manually !) of true energy in and out. Hence I can see the various errors in the Fox cloud recording.

c) I am very aware of the "round trip" efficiency of the battery system. My installer's quote put 80% on this. So when I charge from the grid and then use the battery to power the house, I will consume 10kWh for every 8kWh of "yield" to use the confusing Fox wording. This may explain some of your discrepancies.

d) Finally, your apparent low Solar PV of late actually seems quite good to me (as reported in (b), my figure is about 1.1kWh/day). I don't know where you are in the country but here in the East Midlands, I checked the last few years of solar energy as measured by satellites for this location and I saw roughly 2% of the total annual energy available comes in months of Dec or Jan whilst 14% comes in each of June and July. Of course, the days are shorter and the sun is lower in the sky, so if you are getting 1 to 2 kWh per day on average (with recent cloudy conditions), I expect you will get 7 times that on average in the peak summer which fits with around 5kW peak of solar panel capacity.

Hoping this can add light (and not heat) to the topic and with apologies for any teaching of my grandmother... stuff.

Best Regards

[SolarJunkie]

After spending quite a lot of money on a solar system install, I can fully understand the need for some people to want to interrogate the system to be able to analyse their returns, however, unless you are assessing what to do with any excess generation, or trying to decide on expanding the system for some reason, then the figures obtained, don't really have any bearing on the performance.
If the system is fully operational, it will give whatever return it is able.

Ultimately, you will pay for the electricity that you consume and with a solar install, particularly with batteries, this is significantly less than without.

FWIW, my system generated 9KWh of solar yesterday and 8.5KWh today, so getting around 5KWh from a 5KW system in winter, isn't too shabby in my opinion. To put it into context, that is more than my average daily use and if there was an effective way to automate smart charging, I needn't have used the grid to charge my batteries last night.

[Dave Foster]

I’ll try and pick off as many points as I can, probably the most important one is why do the numbers often not tally and the answer is usually down to system losses, these losses include charging the battery from the grid which requires an ac to dc conversion and then the natural RTE of the batteries (basically chemistry and resistance which generates heat).

Fox quote the HV2600 batteries as having a 97% RTE which is high for LFP batteries, normally its around 95% but if you add conversion losses I would ‘guestimate’ you are probably losing around 10% in the charge / discharge process. There is also the conversion loss between the solar (DC) and grid, and finally the amount of power the inverter needs to stay awake which it takes from the DC feeds (solar / battery), it’s usually about 70w.

It isn’t easy to work all these things out from the Fox statistics, I use a Home Assistant wired into the inverter to give me almost real time statistics, most importantly it allows me to calculate daily system losses - so for example my system today at 6.30pm i’ve had 3kw of solar, my battery has charged off grid 10.94kw, has discharged by 8.84kw and the total system losses for all of that are showing as 2.93kw.

I’m not absolutely sure on this but I think the Total Yield on the site overview monthly stats page is consumption yield plus losses, does Total Yield get close to your Octopus stats ?

On the device page Total yield is everything that has come out of the inverter i.e. all the output the hybrid inverter sends to the house load (either from solar or batts) - you’re right in that you need to take into account both charge and discharge of the batteries from grid or solar to avoid double counting and any solar that passes through the inverter to the load.

I don’t see the labels for charge/discharge energy totals on the web site,I only see Charge and Discharge on the device statistics page? but either way they will be including any charge (or discharge) from solar or grid.

At this time of year when the sun’s elevation is so low, unless you have clear line of sight all around you, you will likely be getting shadows from gables, trees, other houses etc.. and probably only clear solar between 10 and 2.30pm. Yesterday was a very sunny day and my 4.1kw of panels produced 4.6kwh which I thought was quite good, i’m not surprised after battery charge you aren't feeding in, anything at the moment.
Having said that in the summer months (June,July,August,Sept) I would be expecting over 20kw on a clear day so I wouldn’t be disheartened by your output now, from March you should be seeing double what you are now and increasing.

I know it’s frustrating but I would highly recommend looking at a home assistant which is basically a small micro computer such as a Raspberry PI, they are very good at doing real time calculations, providing statistics and lots of other things - i’ve included a screen shot of mine below, it takes all its information from the inverter itself either by connection to the Fox cloud (which gives 5 minute resolution) or by direct connection to the inverter (which gives near real time) - there are quite a few experts on the forums and William Eccles has done some brilliant setup videos to guide you through.

[eepyaich]

As a newbie to FoxESS - just got batteries and in inverter this week for a "no solar" installation (relying on cheap off peak energy to charge the batteries which then power the house for the day) - I'm really grateful for this thread which has definitely helped unpack some of what I'm seeing through the cloud integration. I'm also a newbie getting to grips with Home Assistant, so it's also great to see how folks such as Dave have integrated into the system. Based on this thread, I've already just moved from using the FoxESS Cloud integration to the direct connection.

Dave, thanks for mentioning William Eccles videos - watching one of these reminded me that I had some unused PowerLine network plugs lying around, which did the job to get the inverter connected to my router. I know the instructions for the Home Assistant FoxESS ModBus integration tell you to power off the inverter before connecting the Etherent, but I confess I didn't do this - instead, I used the inverter controls to simply enable DHCP on the Ethernet port which restarted the inverter management unit and seems to have done the trick (much easier than powering off the inverter and battery array!).

The other key thing I've learned from this thread is that the data on the FoxESS Cloud is likely to be a bit inaccurate for various reasons. That certainly matches my experience - I'm seeing the "daily grid consumption" reported on FoxESS cloud to be about 10% lower than my smart meter is reporting. I'll be interested to see if the data from the ModBus is more accurate (although so far, around 2 hours since connecting) the ModBus sensors for the daily accumulated data are still reporting "unknown" (charge/discharge/grid consumption/load).

Dave - your dashboards look great. Any chance of sharing exactly what sensors you are using where (or sharing your dashboard YAML)?

Thanks again to all in the community who are helping us newbies get to grips with all!

[calum]

Best thing to do to learn about HA dashboards etc is to work out what you want to know / be able to check about your own installation. Simple stuff like "how much power did I put into the battery last night?" is useful, but you can also log that night over night, and compare it to the temperature of your battery bank (if you're connecting via RS485 you can see the values the BMS knows about) and then see if the temperature is having an effect on your battery capacity.

Or you can work out how much usable power is coming out of your batteries, and hence what the effective cost of that stored power is - and how much you're saving vs using it directly from the mains.

As you learn how to do new things, you'll find yourself having your own ideas about what you could do, what questions you could answer.

[Dave Foster]

Dave, thanks for mentioning William Eccles videos - watching one of these reminded me that I had some unused PowerLine network plugs lying around, which did the job to get the inverter connected to my router. I know the instructions for the Home Assistant FoxESS ModBus integration tell you to power off the inverter before connecting the Etherent, but I confess I didn't do this - instead, I used the inverter controls to simply enable DHCP on the Ethernet port which restarted the inverter management unit and seems to have done the trick (much easier than powering off the inverter and battery array!).

Yes when the inverter allocates an IP address it restarts Manager so you will see it shutdown, then go through checking etc.. and once complete the IP address should be allocated and active on the ethernet port.

Can I just make one quick recommendation, I know it's more fiddly to get RS485 working but you'll also get control of charge periods, min/max soc and work mode this route so it is very much recommended - the more important reason is that the ethernet port has been shown to cause some inverters to become unstable and freeze up - the latest version of Foxess H1 firmware (Master 1.63/Manager 1.70) has actually removed it's operation as several users have recently found out.

The other key thing I've learned from this thread is that the data on the FoxESS Cloud is likely to be a bit inaccurate for various reasons. That certainly matches my experience - I'm seeing the "daily grid consumption" reported on FoxESS cloud to be about 10% lower than my smart meter is reporting. I'll be interested to see if the data from the ModBus is more accurate (although so far, around 2 hours since connecting) the ModBus sensors for the daily accumulated data are still reporting "unknown" (charge/discharge/grid consumption/load).

The Foxcloud tends to be more unreliable as it's only getting data samples every 5 minutes and so short burst of usage (such as a kettle) can be hidden, but that said if you are using the Inverters normal CT clamp device (most users are) these themselves are inaccurate, particularly at low usage - so you might expect to be within 3-4% of what your smart meter reports. Some users have RS485 meters fitted (Chint/Eastron) rather than CT clamps and they are much more accurate.

Dave - your dashboards look great. Any chance of sharing exactly what sensors you are using where (or sharing your dashboard YAML)?

Thanks again to all in the community who are helping us newbies get to grips with all!

Thanks, it's actually quite an old dashboard that one - home assistant is very addictive and you can "will " spend a lot of time perfecting it.

Most of the sensors there are standard Foxess modbus sensors, and using the standard gauge card and custom button card - with the standard gauge card i've gone to the trouble of defining segments (rather than just red, yellow, green) - it shows a finer resolution and I think looks better, i've posted a sample of a card below to show you it's yaml - but if there are any sensors there you are interested in, or have an idea for them - just post it here and i'll do what I can to help.


This is the gauge code, it uses a template sensor that takes the standard load_power sensor and converts to watts (see below)

Code: Select all

type: gauge
entity: sensor.house_load_w
name: House Load
min: 0
needle: true
segments:
  - from: 0
    color: rgba( 0,225,  0,1)
  - from: 500
    color: rgba( 0,215,  0,1)
  - from: 1000
    color: rgba( 0,205,  0,1)
  - from: 1500
    color: rgba( 0,195,  0,1)
  - from: 2000
    color: rgba( 0,185,  0,1)
  - from: 2500
    color: rgba( 0,175,  0,1)
  - from: 3000
    color: rgba( 0,165,  0,1)
  - from: 3500
    color: rgba( 0,155,  0,1)
  - from: 4000
    color: rgba(255,185,  0,1)
  - from: 4500
    color: rgba(255,175,  0,1)
  - from: 5000
    color: rgba(255,165,  0,1)
  - from: 5500
    color: rgba(255,155,  0,1)
  - from: 6000
    color: rgba(255,145,  0,1)
  - from: 6500
    color: rgba(255,135,  0,1)
  - from: 7000
    color: rgba(255,125,  0,1)
  - from: 7500
    color: rgba(255,115,  0,1)
  - from: 8000
    color: rgba(255,105,  0,1)
  - from: 8500
    color: rgba(205, 0, 70,1)
  - from: 9000
    color: rgba(215, 0, 60,1)
  - from: 9500
    color: rgba(225, 0, 50,1)
  - from: 10000
    color: rgba(235, 0, 40,1)
  - from: 10500
    color: rgba(245, 0, 30,1)
  - from: 11000
    color: rgba(250, 0, 20,1)
  - from: 11500
    color: rgba(255, 0, 10,1)
theme: AMOLED Blue
max: 12000
unit: w

And this is the template sensor that does the conversion -

Code: Select all

      - name: "house load w"
        unit_of_measurement: "W"
        state: >
          {% if (states("sensor.load_power")|float(default=0) * 1000) <=1 %}
            {% set hlw = (states("sensor.house_load_w")|int(0)) %}
            {{ hlw|round(0) }}         
          {% else %}
            {{ (states('sensor.load_power')|float(default=0) * 1000)|round(0) }}
          {% endif %}

[Golfchristopher831]

Hi the formula at the start of this thread is ridiculous - what is wrong with the FOXESS developers - surely they can give us what the solar generates. The formula previously was :

Solar generated = (battery charge - battery discharge - consumption yield + Load yield + Feed in yield)

When I speak to my Solar company they try to claim that Total Yield is the Solar , but this is clearly wrong as it has a positive number at night.


My suggestion is that we leave feedback on the appstore to try to get this resolved.


Chris

[Dave Foster]

Hi the formula at the start of this thread is ridiculous - what is wrong with the FOXESS developers - surely they can give us what the solar generates. The formula previously was :

Solar generated = (battery charge - battery discharge - consumption yield + Load yield + Feed in yield)

When I speak to my Solar company they try to claim that Total Yield is the Solar , but this is clearly wrong as it has a positive number at night.

Chris

Hi Chris, I don't think you will find many on here that would argue differently - If you have a solar only inverter Total Yield is always solar generated, but with a hybrid it also includes the power provided by the batteries - in effect it is inverter total output, not solar yield.

The better news is that if you download either the Energy Stats app, or the FoxESS V2 app they provide a solar yield sum