Who here has the best dump load management system going on? (revisited)

Posting attention people who participated in an earlier thread, and others who might have valid input.
@SeaGal @efficientPV @kmin @AntronX @sunshine_eggo @rogerdw @kommando @Mattb4 @AlaskanNoob @Hedges @Michelle Konzack @ThaiTaffy @Supervstech

Load dump management is an important topic that has been raised multiple times. In researching it, it seems the best way to make use of excess power from PV is using the methods outlined by @efficientPV and @kmin, described as follows:

"You need a controller that looks at array voltage. When array voltage goes over Vmpp, you start to divert power till the voltage drops back to Vmpp."

See related posts here and here.

Seems like few people on this forum actually have this working. @efficientPV seems to have it worked out but it's unclear what his design entails.

Does anyone on the forum have this working that could offer full details on their method? Specifically a method using Vmpp monitoring because what we want to know is when the PV array is capable of sending more power but isn't. i.e. PV generation potential exceeds the load (and the battery is already full). Having said that, if you have an SoC method working great with something like Home Assistant, open to that as well.

Please refrain from posting theoretical solutions to keep this thread on topic. This is specifically for people who have actually gotten this working in their own environment. Or if you're interested in being part of a group trying to come up with a solution to this problem please PM me. Thank you!

Thank you @AlaskanNoob, that was a great post you made on this topic. Did you manage to get something working?

For starters, in this post, @efficientPV said that this is the controller he is using. Multiple people asked if he could specify which controller he was using and AFAIK he refrained from doing so, instead said there are "lots of ways you could do it." Can anyone identify this controller?

As @AlaskanNoob said, massive amounts of potential solar power is lost because there's no easy way to use it / divert it. Maybe as a group we can figure out a simple, inexpensive way to make that wasted power more readily accessible to the solar power community. Anyone interested please PM me.

kolek said:

For starters, in this post, @efficientPV said that this is the controller he is using. Multiple people asked if he could specify which controller he was using and AFAIK he refrained from doing so, instead said there are "lots of ways you could do it." Can anyone identify this controller?

View attachment 231621

Click to expand...

He built it himself. Schematic capture, PCB design and layout tools, and custom short-run PCB manufacture is pretty straightforward nowadays.

I use SOC, as it's easy to harvest from the inverter in closed-loop mode, but there are other ways. I use an RS-485 interface and some custom Python code, but it's not rocket science. I have a database table with info about https://www.shelly.com/en-us/products/shop/shelly-pro-1pm interfaces (one per HVAC Split, a few others for sump pumps and such). Two of the entries are SOC_on and SOC_off. Every minute I go through all the devices and:
[...]
if ((RedisState == 'Off') and (SOC >= int(Dict['SOC_on']))):
print("Turn it on!")
ShellyControl(Dict,OutputNum,True) # make the call to turn it on
RedisDB.set(RedisName,"On")
MailMe("Turned {} {} On".
format(Dict['Name'],Dict['OutputName{}'.format(OutputNum)]),
"SOC: {}% SOC_On: {}%".format(SOC,Dict['SOC_on']),
False)
if ((RedisState == 'On') and (SOC <= int(Dict['SOC_off']))):
print("Turn it off!")
ShellyControl(Dict,OutputNum,False) # make the call to turn it off
RedisDB.set(RedisName,"Off")
MailMe("Turned {} {} Off".
format(Dict['Name'],Dict['OutputName{}'.format(OutputNum)]),
"SOC: {}% SOC_Off: {}%".format(SOC,Dict['SOC_off']),
False)

Because it's all DIY, and there are a lot of interleaved parts, there's no sense in saying exactly what I did, as no-one else is going to want to clone my setup exactly but this has gotten my SOC graph over the last few days to look like:

This has meant my 'discretionary' loads used 44.2 KWHR yesterday. I'm not fussed about using every single spare KWHR that I _could_ have harvested, because 'dump' loads tend to become 'discretionary' -> 'normal' -> 'important' -> 'critical' and then I'm doing the #GiveAMouseACookie thing and adding more panels, batteries, and loads just to keep up, and then I need a bigger generator for those cloudy weeks.
My batteries turn the corner between discharging and charging around 7:30, they are fully charged around noon, and then they equalize/balance for a few hours before picking up the load at sunset.

I'm not thrilled with string voltage as a metric for 'wasted power', while it's true that Vstring>Vmp tells you moment by moment if you are missing potentially harvestable energy, it feels like you are adding another layer of MPPT on top of your existing SCC, and the signal for turning your dump load on and off is too noisy:

I'm not even sure where I'd set the threshold for each of my six strings. SOC changes much more slowly, so my discretionary loads run for hours at a time instead of minutes.

The best dump load would be the grid, and most grid-tie and AIO inverters send their surplus power there, where the capacity is infinite and someone else can use it at the moment. Of course, Net Metering is going the way of the dinosaur, and the PoCos are trying the remove all the economic incentive from it (aka capture the profit for themselves).

I really like the Victron 150/60 and up with their relay that can be programmed for "float status" ... Imagine what you could do with that

kolek said:

Thank you @AlaskanNoob, that was a great post you made on this topic. Did you manage to get something working?

Click to expand...

Not yet. We don't even have our other five arrays up yet.

kolek said:

For starters, in this post, @efficientPV said that this is the controller he is using. Multiple people asked if he could specify which controller he was using and AFAIK he refrained from doing so, instead said there are "lots of ways you could do it." Can anyone identify this controller?

View attachment 231621

Click to expand...

This was a failed venture years ago to get a friend's son into a little money making venture. I figured he could sell two a month on ebay. I bought all the parts to make more than a hundred of them. I built a number of pre production boards and provided them for evaluation. I only heard back from one. One made me send it FEDEX and it still has never been installed. With no interest, my friend's son lost interest too. A production version was never built. And his situation has also changed, Making 35K a year and still in high school. I understand now it is just too technical for most and I don't want to be involved with bad installations that are dangerous. These can be expensive if not ordering parts in quantity when only making a single unit. Versions of this design have been built around the world and they work well. This diversion method for best effect does require that the array have a natural power point, the controller will adjust for temperature. If someone is interested in laying out a board, I'll help with design and wish them luck. It just won't be me selling boards. Would like to find someone in Europe where there is more interest in heating water.

efficientPV said:

This was a failed venture years ago to get a friend's son into a little money making venture.

Click to expand...

That is fascinating that you had this as a concept for a venture.

Marketing is really at the core of any venture like that. If a company like Victron made it, I'd think they could sell a million.

Really appreciate you making that information available. Not sure how many potentially interested parties will see this particular thread and read your post, but I think it's of interest to many on this forum.

wpns said:

I'm not thrilled with string voltage as a metric for 'wasted power', while it's true that Vstring>Vmp tells you moment by moment if you are missing potentially harvestable energy, it feels like you are adding another layer of MPPT on top of your existing SCC, and the signal for turning your dump load on and off is too noisy:

Click to expand...

Wow that is a good point. I didn't even think about the noise.

I see that clearly in our install. Japan has very few cloudless days, so every 30 seconds the sun is obstructed and as a result my string voltage is whipping all over the place, output goes from 5A to 130A and back to 5A in the space of 15 seconds and sometimes this goes on all day. So maybe SoC is a better metric to use afterall?

AlaskanNoob said:

Not yet. We don't even have our other five arrays up yet.

Click to expand...

The rotating array mounts you have are so nicely done, I encourage anyone who hasn't seen his build thread to check them out.

wpns said:

Net Metering is going the way of the dinosaur, and the PoCos are trying the remove all the economic incentive from it (aka capture the profit for themselves)

Click to expand...

That describes exactly what's happening in Japan.

@nx6 or @Daddy Tanuki would know more about this than I do, but I believe in Japan if you have more than 10kw worth of panels, you are required by law to sell all your power to the power company, who's pays you a fraction of the market price for it.

The amount they pay goes down continuously, while the price they charge for electricity goes up practically every month.

kolek said:

That describes exactly what's happening in Japan.

@nx6 or @Daddy Tanuki would know more about this than I do, but I believe in Japan if you have more than 10kw worth of panels, you are required by law to sell all your power to the power company, who's pays you a fraction of the market price for it.

The amount they pay goes down continuously, while the price they charge for electricity goes up practically every month.

Click to expand...

yep. our home solar install which is a grid tie and was installed by one of tepco's "authorized installers" (read that as cronyism), just finished paying for itself and our contract unlocks next year. when it does our payback will readjust to whatever the current low level is. Originally when we bought the system to incentivise installs they were locking in your payback for a period of time, due to the size of our home system 8kWh of panels ours was for I think 12 years. they have more than paid for themselves in the lower monthly electric bill and the monthly payout of excess, but once they drop the rates it will not be worth selling to them. better to build a battery bank and power yourself through the night.

I am considering unhooking it but setting it up as the primary with the grid as a backup. it would probably be a pair of the tristar 600 volt scc into a 140kWh lithium bank inverters TBD. the tristars could wire right up to the current panels with no changes and go from there.

Clearly there is a large misunderstanding of this concept. Diversion is done in hundredths of a second. This catches every change in a cloud, far faster than a MPPT. It isn't really an added layer. wpns already has mor of everything than is needed. Diversion would have no value unless he had a swimming pool to heat. Large strings usually have some shade issues that make it impossible to have any set point. These diversion products aren't available because only a rare few have the understanding to use them effectively.