Scratch built solar PV divertor

[Radian]

2 hours ago, SteamyTea said:

So you are boiling enough to make 16 mugs.

No, because we don't take 237ml of milk with our 63ml of tea. What are you talking about? 😵

[MikeSharp01]

On 22/06/2022 at 14:17, Radian said:

building a list of places where I could prioritise the diversion of excess energy so I'm giving everything a chance.

So how do you propose to control the dispersion to the various devices and deal with the heat store / destination not needing any more EG the kettle boiling and tripping out / whistling the place down. 

 

I suppose you might use the joule bucket, if it can resolve cycle by cycle, to verify although you can monitor the current draw of the kettle. IE you have prioritised the kettle but the joule bucket is still growing so it is not taking the juice so move down the priority list perhaps then work out an algorithm to creep back up the priority list, might waste a cycle for each destination not up for juice.  

[SteamyTea]

With PV, you could take a statistical approach.

If it has produced for say 10 seconds, then you can look at the probability it will produce for the next 10 seconds.

Then, after a minute, take that probability and use it to predict the next minute, while still gathering data at the 10 second interval.

You could make it posh and start to predict power due to time of year i.e angles and altitude, length of day.

It would not respond to variation as such, but may allow import limiting i.e. disconnect heavy loads from grid while still allow them to use PV power. Some kind of signal controlled grid isolator on, say, the immersion heater would be needed.  This would allow other loads to draw from the grid while the PV is dedicated to one task. While also allowing any excess loads i.e.vacuum cleaner or kettle, to be still powered by excess PV i.e. greater that 2.8 kW 

Edited June 26, 2022 by SteamyTea

[Radian]

1 hour ago, MikeSharp01 said:

So how do you propose to control the dispersion to the various devices and deal with the heat store / destination not needing any more EG the kettle boiling and tripping out / whistling the place down. 

 

I suppose you might use the joule bucket, if it can resolve cycle by cycle, to verify although you can monitor the current draw of the kettle. IE you have prioritised the kettle but the joule bucket is still growing so it is not taking the juice so move down the priority list perhaps then work out an algorithm to creep back up the priority list, might waste a cycle for each destination not up for juice.  

 

Because we have a very rigid routine, a boiled kettle is expected at certain times of day. So if we always leave it primed with cold water and if within, say, 15minutes of the usual cuppa demand then gate through any surplus as a priority over other loads. This would only be helpful on days with variable PV and this is where SteamyTea's predictive algorithm might come into play. The solcast api is ideal for this with 50 free requests per day. I'm also prepared to let the kettle driver notify me if boiled earlier than expected. The regular cutout  and CT will see to this. To disable it we would simply leave the kettle switched off.

 

In the absence of battery storage, the aim is to advance/retard certain loads that can be time-shifted to make hay while the sun is shining. It will only be worth the trouble if there are lots of examples like this.

[SteamyTea]

36 minutes ago, Radian said:

In the absence of battery storage, the aim is to advance/retard certain loads that can be time-shifted to make hay while the sun is shining.

Kills the theoretical efficiency because thermal losses, that are unused, become parasitic loads.

 

With your baseload, especially from the IT kits you have dispersed over the house, you would be better off putting them all in one insulated box, then making a small heat pump to scavenge the energy and put it into water.

[Radian]

51 minutes ago, SteamyTea said:

Kills the theoretical efficiency because thermal losses, that are unused, become parasitic loads.

I'm not talking about massive time-shifting here. Not like the stuff will be on when the house is empty!

 

OK, here's another example you'll love: Tumble dryer.

 

So yes you'd be right to point out the folly but you've no doubt heard the all-important fluff-effect. So let's say the line dryer has done its bit and the crispy clothes are brought back inside and placed inside the dryer - with the Sun going in and out (very much like today). A small intervention with the OEM pause button has allowed me to enable/disable the dryer as the PV picks up and excess becomes available. There's no hurry for the results and a notification for when the job is done (or not).

[SteamyTea]

17 minutes ago, Radian said:

small intervention with the OEM pause button has allowed me to enable/disable the dryer as the PV picks up and excess becomes available. There's no hurry for the results and a notification for when the job is done

Surely this is so inefficient, basically heating up an element and a few grams of air, to have it cool down again, just to have it warm up again.

Or are you just running the tumbler without a temperature rise?

You get used to toweling off with something hard and raspy.

 

[Radian]

The thing is, some people in this household would just slam it on whenever they liked. If the power wasn't coming from the sun they'd do it anyway. I can ask nicely to wait until there are optimum conditions ~ or hot-melt glue an ESP8266 into the machine and leave it to that. The 3V3 is already there, the pause button just needs pulling to 0V and bob's your uncle. This is on a slow control (over MQTT) when the Joule buffer is filling at a similar rate to the 1.5kW heating element that's allowed to empty it.

 

You also have to allow for the fact that I've got a free part-reel of these WiFi chips and relish the opportunity to make my own smart appliances for nothing other than the time I enjoy spending on coding them. Anything with spare 3V3 available is fair game to me if it draws more than a few hundred watts and can be time-shifted.

[ProDave]

2 hours ago, Radian said:

So yes you'd be right to point out the folly but you've no doubt heard the all-important fluff-effect. So let's say the line dryer has done its bit and the crispy clothes are brought back inside and placed inside the dryer

Interesting. I find giving the wet towels a spin in the fluff making machine just for half an hour, which is not enough to dry them, and then completing the drying naturally works to give the desired fluffyness.  Interesting that ti also works the other way, as long as you don't let them dry completely naturally otherwise you get hard towels, and complaints.

[SteamyTea]

7 minutes ago, ProDave said:

Interesting. I find giving the wet towels a spin in the fluff making machine just for half an hour,

We should do a 'citizen science' project.

Find out the best way to fluff a towel.

[ProDave]

27 minutes ago, SteamyTea said:

We should do a 'citizen science' project.

Find out the best way to fluff a towel.

When @Radian has designed the "fluff sensor" that stops the TD as soon as the required fluffyness has been reached, even though it will not be "dry" he will have a worthwhile product.

[MikeSharp01]

This is a great example of the 'socio-technological gap' technology has the capability to do things that society finds difficult. So the entrepreneur needs to path the gap closed for the society, show them the way and if you are lucky they will go for it and allow load shifted fluffing. If you are an Economy 7 customer you should already be doing it anyway.

 

As with any problem / opportunity, your world view matters, and you can either go through it, round it or over it. (Others may see this differently EG fix it upstream, midstream or downstream ) So in this case we might spend our energies, pardon the pun, in three broad areas (from a big picture point of view and there are many other ways to see this).

 

1. Through it / midstream - we are already doing that by showing how things might be done the way we are doing it already but slightly better. Conserving energy by load shifting and educating the populace. [Evolutionary?]

2. Round it / downstream - Seek a. to educate people that fluffy towels are a thing of the past [Evolutionary]or b. make it so towels fluff themselves - which might also be an example of upstream as well but hey bear with. [Perhaps revolutionary as many other things would get and stay fluffy]

3. Over it / upstream - Fight to get a fair price for our exported energy so we can use the communal supply as and when - fluff whenever we need.

 

In 1 and 3 fluffing is still with us while in 2 we no longer need to worry about it. But 3 would have massive additional opportunities. I guess the question then is who will be the biggest winners. (Don't worry either way it won't be you or I)

[SteamyTea]

9 minutes ago, MikeSharp01 said:

Conserving energy by load shifting

Load shift can also create situations that increase usage i.e. heat a store many hours before it is needed.

10 minutes ago, MikeSharp01 said:

educating the populace.

The RE industry is having a private chuckle at the moment as they have been banging on for over a generation (2 or 3 generations in some cases) about usage.

If, after 60 years, the message has not gone in, we need to have high prices to change behaviour.  Education does not always work.

[ProDave]

On 15/06/2022 at 12:45, Radian said:

LTSpice - wonderful bit of free software from Linear Technologies. It simulates circuits to save you blowing things up for real!

I must have a look at that.  Last time I used Spice, it was a command line interface, and it was quite a task to integrate it into ORCAD (which I still use) to run a simulation direct from the schematic entry tool.

 

I see they don't have a Linux version, I will have to see if the windows version runs under WINE.

[MikeSharp01]

17 minutes ago, SteamyTea said:

RE

Sorry brain fog RE?

[MikeSharp01]

21 minutes ago, SteamyTea said:

Education does not always work.

Yes it always does but the outcomes are not always predictable so it may appear not to. If you have an education you are educated if not you are not. Neither way round precludes genius or wastrel. 

[ProDave]

4 minutes ago, MikeSharp01 said:

Sorry brain fog RE?

Renewable Energy?

[SteamyTea]

7 minutes ago, MikeSharp01 said:

but the outcomes are not always predictable

Yes, I will accept that.

But if the outcome is the opposite of what you intended, then it has failled.

 

For decades, university economics taught that people bought rationally.  This explained the classic supply and demand charts.

2007/8 showed that this is not the case and people can act quite irrationally in a free market.  I am glad they do, we are better off because of it.

[MikeSharp01]

29 minutes ago, ProDave said:

Renewable Energy

Thanks - why didnt I get that.

 

Not sure it's in their interests to get us to cut back they want to sell more product not less but they control demand with cost, in their case the cost of making what they make using gas - daft but seems true. 

 

P.S. we are way off topic here do we need a new thread.

[SteamyTea]

3 minutes ago, MikeSharp01 said:

Not sure it's in their interests to get us to cut back they want to sell more product not less

They are not really bothered about how much we buy, it is the profitability that is important.  That is controlled though marginal costs.  So they would be happy to supply 998 MW from a single 1 GW generator than 1002 MW from 2 1 GW generators.

They do the exact same calculations that we do when trying to work out if PV is financially valid, except we try to self consume, rather than sell on the open market.

[MikeSharp01]

30 minutes ago, SteamyTea said:

But if the outcome is the opposite of what you intended, then it has failled.

No you have overlooked the fundamental reason for education.

 

Education is about endevoring to take closed minds and make them open. The basic goal is founded on the principle, in a free country such as this notionally is, that we want all members of society to be able to either agree with each other, have the wherewithal to change the minds of others or find common ground on which both can agree. It really is that simple.

 

Sometimes it does not equip people well to see the other side of the argument or even be eloquent in their discourse but this does not mean either side is in the wrong. It mearly means that one or both sides need to increase the reach and strength of their argument(s) NOT the volume of their voices.

[SteamyTea]

31 minutes ago, MikeSharp01 said:

Education is about endevoring to take closed minds and make them open

Or passing the OFSTED at one of the places I worked at.

 

[MikeSharp01]

15 minutes ago, SteamyTea said:

Or passing the OFSTED at one of the places I worked at.

Been there, done that, and now OFSTED have a remit in universities via the degree apprenticeship, need to keep an eye again. However the basic remit remains the same take closed minds and open them. 

[Radian]

So I’ve been running my diverter for just over a month now and I’m ‘leaking’ a little bit of power to the grid.
I have taken a leaf out of @SteamyTea's book and further refined the calibration of my power measurements by adding an optical sensor to time the pulses from my Elster AS300P (4000 imp/kWh) so I’m quite confident that my power measurements are decent enough. I'd show a photo of it but its an awful bodge of masking tape holding the phototransitor and wires on the face of the meter.

 

What seems to be happening is that if power is only just tipping the balance by a few 10’s of Watts of export, and I sit and watch the electricity meter while the Joule bucket is slowly filling and emptying (as witnessed by watching my Joule bucket animation) the bucket fills and empties cyclically for several 10’s of minutes - until the flashing LED on the meter suddenly goes on hard and the little tamper symbol comes up on the display. This evidently signals that power is going into the grid...

 

 

I have now added another debug bitmap to my ESP32 firmware to see what’s going on in more detail on these occasions:

 

 

• Each of the 256 horizontal pixels is a 20ms mains cycle (5.12s for the full image)
• The Red line marks the (assumed) 3600J limit to the Joule bucket.
• The Green trace is real power with zero half way up the image (as is cyan, but filtered)
• The yellow lines are the upper and lower Joule limits of my dump algorithm (temporarily set low for investigation into the leakage)
• The dark blue trace is the Joule bucket content as tracked by my software algorithm.
• The grey verticals represent cycles when the Elster meter LED is on.

So in this example, my algorithm thought we were comfortably operating in the ‘unchallenging’ 500J to 1000J range yet after a while of this cycling (maybe half an hour) the anti-tamper symbol and LED come on. Clearly dead-reckoning the state of the Joule Bucket is bound to lead to inaccuracies over extended periods of cycling. I have not seen any prior discussion of this, but it seems inevitable to me.

As a resolution, the only absolute measure of the Joule bucket state is when the anti-tamper/LED shows up so I will max up the bucket content when this is detected. Likewise at the other limit, a deliberate periodic drain-down of >3600J (1Wh) every so often would improve accuracy at the expense of a little unnecessary electricity consumption. Can anyone tell me if they think I’m on the right track here?

[ProDave]

Fascinating analysis but the amount of leak is tiny in real terms.

 

How are you managing to dump all your generation?  I have exported 1.8kWh this week, almost all around mid day if nothing else is on in the house and the PV generates more than the immersion heater can absorb, and it's too hot to have the additional 700W convector as a dump load (unless out of spite I put it outside)  Is it even worth looking to see if I can get an immersion heater that uses more than the 2.8kW mine seems to max out at?