Suggestion Adding the daily temp curve

Is it possible to add inverter temperature and GridBoss temperature also?

How do they even get those estimations without any input as to what solar PV inputs we have?  Seems like that data would be helpful in a more accurate estimate.

Not to mention having a way to include AC coupled input into the "solar" generation equations. 

@dougl 

I believe they are using the forecasted cloud cover in a "sunshine" formula (i have not a clue). If they already know the forecasted temp a "shade" factor based on predicted cloud cover seems "logical" at least.  Maybe we'll get lucky and have someone in the know tell us the actual methodology.

 This also appears to provide a predictive trace, as it  appears to adjust as the day goes along.

@dfwtinker I agree but the number of kWh they are predicting has to be estimated by a guess of what size our arrays are. 

I would think they could logically get far better results if they knew at minimum the MPPT connected array sizes. If they were using historical data then they would likely show the notch in the morning hours where I have shading. Instead they estimate a smooth curve so it's really not trustworthy and I can probbly do better looking at the projected cloud cover myself.

I wondered the same thing. Perhaps they should allow a system setting of a "prediction multiplier" that uses the historical data. In my case, because I have a 12.8KW array, I (obviously) always bet the prediction on sunny days but (surprising to me since mine are bi-facial) almost aways under perform -vs- the prediction on rainy/mostly cloudy days.

Like others, I do still appreciate the production prediction curve and estimates.

Posted by: @dougl

↑

@dfwtinker I agree but the number of kWh they are predicting has to be estimated by a guess of what size our arrays are. 

I would think they could logically get far better results if they knew at minimum the MPPT connected array sizes. If they were using historical data then they would likely show the notch in the morning hours where I have shading. Instead they estimate a smooth curve so it's really not trustworthy and I can probbly do better looking at the projected cloud cover myself.

Good Morning @dougl 

I totally agree that you would get far better results.  However adding that single variable complicates the behind the curtain effort needed to make that work from a programming and system engineering perspective.   Whereas with the addition of a single curve based on your zip code (which the app already has);  Adding in a Customer introduced variable gets much harder.

Now you have to create, store and update tables for each product you choose to support.  Then you have to create the bulletproof input logic to accept the choice.  Then you have to worry that the calculations are exactly the same or do they change because the different inverters supported do things differently (internally).   And ....    The list of questions and considerations needed just for the single option explodes, and profit vs free utility loses every time.   But Yeah! it would be nice!  

Maybe the EG4 engineers will surprise us and add it to something they already have in the works . . . 🍻 

p.s. Before anyone pounds on me for my lack of detail, it is intentional, I'm trying to follow the KISS methodology.  I assure you that as an Enterprise Arch. managing  Metrics & Observability and 45+ years in IT; that I can go into great detail if you need me to.

@dfwtinker well they do store lots of stuff from our inverter and details about the wifi, network, etc. They render the curve in the browser(Javascript ? ) so pulling in a kW system size variable and scaling the graph was all I was asking. But otherwise I don't feel like I can trust it other than a very vague idea if it'll be sunny or not but my local weather chart data also tells me that with an icon.

No biggie, was just a thought that there might be a way to make it more useful and relevant.

@dougl 

I hear ya!  I have been downloading some of the monthly data and re-uploading it to a NewRelic metrics & observability account i maintain to keep stats on my two home machines and my synology NAS.   That way I can create some custom metrics. One of the data points I have not been able to find a source for is the minute by minute (5 min would be ok too) data feed.   I am beginning to think that I will have to install my own weather station and upload that data to NewRelic.  heavy sigh.. 

I hope they come up with something we can use as well, i'm sure it will be "pot-luck".  Hell I'm even willing to pay; as long as it is reasonable.  Storage is cheap these days, If I can afford a measly 100TB triple redundant NAS for my photography, they ought to be able to work out something reasonable for a growing business concern (no I'm not wealthy, I have two daughters and 4 grand-daughters; how can I be?  lol). 

Priorities, demand and money will seal the deal and make it happen.  or NOT!  

L8r

Here's an idea which might help make the prediction data useful, add Yesterday's actual data curve.  We see "Today Solar Energy" prediction getting updated as the day passes so wouldn't it be handy to see the actual previous day's curve with the predictions. instead of having to switch to data mode and see the data?

@DougL

OH 😆 I like they way you think! 

That is not a new external data request/query/download, but reusing the data they already have, the extra space to hold that info for an extra day should not amount to more than 1 or 2 KB max.  Plus the new space required would not grow after initial deployment as the same space is being repopulated every day.

L8r

Posted by: @dfwtinker

↑

@DougL

OH 😆 I like they way you think! 

That is not a new external data request/query/download, but reusing the data they already have, the extra space to hold that info for an extra day should not amount to more than 1 or 2 KB max.  Plus the new space required would not grow after initial deployment as the same space is being repopulated every day.

L8r

And the human brain can work with trends and generalizations very well. We will know what the previous day's weather was like, what the current weather is like and by seeing the previous day's actual generation curve along with the predictions we can extrapolate how much we can expect.

This has become more important to me since we switched to off-grid and stopped selling for peanuts back to the grid from our house battery. Instead we are now charging 2 neighbor's EVs with excess solar generation on really sunny days and from our battery. One neighbor needs 30kWh for a trip to work tomorrow but our house battery only has ~12kWh to spare due to a few cloudy days and charging our own 2 EVs.  Not knowing well what the predictions will mean causes me to have to periodically check out house battery to see if we need to stop charging her EV early to be sure we have 2 days of charge storage for our house.