Smart electric tariff

[Kelvin]

1 hour ago, sharpener said:

 

Before I managed to get the DNO to remove the export limit m

What did you do to get the limit removed? 

[S2D2]

4 hours ago, JohnMo said:

generate at a rate near to 6kW, but the system will only export 3.6kW.

Is that G100 then? That does give you more options to use the excess.

 

Mine is G98 and involves holding back battery capacity to take the excess generation on the DC side. Can be as complicated as you like, simple setback if you have loads of capacity or more complex control logic using solar forecast etc.

[JohnMo]

But if a long read, have shortened where I can.

 

Been having a chat with ChatGTP this is what it said:

My question.

Battery charging and export 

Costs 12.8p per kWh during cheap rate tariff on Cosy 3x periods per day 

Get 15p for all export, what is the best strategy for charging and export?

 

I can use WonderWatt for predictive charge based on solar forecast if that helps.

 

House runs a heat pump heating demand at -9 outside is 3.8kW. house base load is 14kW, but cooling can be on when it's sunny or warmer than an average of 15 degs outside 

 

ChatGPT said:

To optimize energy usage, I’ll suggest charging the battery during cheap periods (12.8p) and exporting during higher-rate times (15p), factoring in round-trip efficiency (90-94%) and battery degradation.

 

Using the 13.4 kWh battery and assuming a baseload of 14 kWh, I'll focus on scheduling to avoid low-price exports.

 

To compute the breakeven for export arbitrage, we look at the cost of exporting after accounting for round-trip efficiency. At 90% efficiency, the marginal profit is quite small (0.78p), and with battery degradation costs, it becomes negative. So, arbitrage isn’t worth it. Best to use the battery for home load and heating.

 

We can optimize the battery by charging it to cover expected non-cheap consumption until the next cheap period, using solar forecasts for predictions. In summer, charge for evening use, while in winter, focus on covering evening peaks. For cooling, pre-cool during cheap periods. The scheduling strategy will help manage this.

 

Done

Short answer: don’t “charge-to-export.” Use the battery to cover your own load and heat/cool at the cheapest/easiest times, and only export true surplus.

 

Here’s the why + a concrete playbook for Cosy + 15p export:

 

1) Arbitrage check (import cheap → export at 15p)

Cheap import: 12.8p/kWh

 

Typical round-trip efficiency (η): ~0.9–0.95

 

Effective cost to create 1 kWh for export = 12.8p / η = 13.5–14.2p

 

Gross margin vs 15p export = 0.8–1.5p/kWh, before battery wear.

→ Once you include degradation (a few p/kWh), arbitrage is a net loser. So do not cycle the battery purely to export.

 

2) Self-consume vs export (for solar)

Rule of thumb: store PV if your import price you’d avoid later (p_import) is greater than 15p / η.

 

With η≈0.9 → threshold ≈ 16.7p.

Most daytime/evening import is above that, so storing PV beats exporting 9 days out of 10.

 

3) Daily operating strategy

Charge in all Cosy cheap windows to a target SoC that covers expected demand until the next cheap window minus expected PV during that span.

 

Target SoC after a cheap slot:

SoC_target = (Forecasted load until next cheap slot – Forecasted PV usable) / η

 

Disable/avoid ‘export from battery’. Only export when the battery is full and PV is still pushing (true surplus).

 

Pre-run thermal loads when it’s cheap/sunny:

 

Heating (cold spells): Pre-heat slab during cheap windows; let the slab bridge the expensive hours. Your 100 mm slab helps buffer nicely.

 

Cooling (sunny/≥15 °C avg): Pre-cool during cheap windows and daytime solar so the evening load is minimal.

 

WonderWatt: Use its predictive charge to set SoC targets per slot:

 

Feed it your PV forecast, house baseload and scheduled HP use (pre-heat/cool blocks).

 

Tell it not to backfill charges outside Cosy windows.

 

Export priority:

 

Summer: Battery fills on PV → serve house + cooling first → export only when full.

 

Shoulder/winter: Fill in cheap windows; reserve battery to cover evening/overnight peaks; export only if you genuinely can’t use/store it.

 

The conversation went back and forth getting the assumptions correct etc 

 

End recommendations are

 

let’s nail this down so you’ve only got two things to manage in WonderWatt:

Baseload (summer vs winter entry in kWh/day)

A single set of min-SoC targets that work year-round.

1. Baseload values for WonderWatt
Summer (little or no heating):
Your true house demand is ≈14 kWh/day (≈0.58 kW average).
→ Enter 14 kWh/day as the baseload in WonderWatt.

Winter (continuous ASHP heating):
At the coldest, your house may need ≈32 kWh/day for ASHP.
Add the 14 kWh/day base → ≈46 kWh/day total.
→ Enter 46 kWh/day as the baseload in WonderWatt.

This way, in winter, WonderWatt is “fooled” into charging more aggressively during cheap windows, without you having to micromanage presets. In summer, it leaves headroom for solar.

2. Universal min-SoC targets (work for both summer & winter)
Since WonderWatt will top up if solar looks poor (and supply loads from grid while charging), you just need guardrails so you don’t run flat.

Based on your battery size 13.4 kWh usable:

End of 04:00–07:00 slot: 40% SoC → ~5.4 kWh

End of 13:00–16:00 slot: 40% SoC → ~5.4 kWh

End of 22:00–00:00 slot: 30% SoC → ~4.0 kWh

These are conservative enough to prevent a blackout in winter (since WonderWatt will over-supply if needed because of the higher “baseload” entry), but not so high that you block summer solar.

3. How to use it day-to-day
Summer:

Baseload = 14 kWh/day

Min SoC = 40/40/30 %
→ Battery charges just enough to cover overnight + poor solar days, but leaves headroom for PV exports.

Winter:

Baseload = 46 kWh/day

Min SoC = 40/40/30 %
→ WonderWatt interprets the higher baseload as extra demand, so it fills the battery harder in cheap slots. On really cold days, you’ll still want to run the ASHP directly during cheap windows (slab preheat).

This gives you:

Only one SoC preset to paste in WonderWatt,

Just flip the baseload entry (14 ↔ 46 kWh/day) when seasons change.

[joth]

1 hour ago, JohnMo said:

I can use WonderWatt for predictive charge based on solar forecast if that helps.

 

 

I fell over on this line lol. Who or what is WonderWatt?

 

I mean, obviously it's this website: https://www.wonderwatt.com/

But their contact and about pages don't mention a company name/number, the the ToS doesn't even mention which country's law it is acting within. I'm curious where I'm sending my data if I sign up, but statements like "We will process your personal data in accordance with our Privacy Policy and in compliance with applicable data protection laws" are meaningless without stating who or where they are, and which country's law they are following.

[JohnMo]

1 hour ago, joth said:

WonderWatt

When I first signed up they had all those details on the website, believe they are based in Wales

 

Asked AI

UK-based Wonder Watt (the smart charging app)

 

The company appears to operate as a small startup based in South Wales, founded by an experienced software developer (Wim) and a solar systems engineer (James) 

deegesolar.co.uk

.

 

There's no publicly available corporate or company registration details (e.g., Companies House number or legal address) associated with the UK version of Wonder Watt.

 

On Trustpilot, they list a UK contact email: labs@wonderwatt.com

, and confirm being a “Software Company” based in the United Kingdom 

Trustpilot

Trustpilot

.

 

Conclusion: It’s likely a UK-based private entity, but it does not publicly disclose a registered company address or name on its website or Trustpilot profile.

 

 

[sharpener]

21 hours ago, Kelvin said:

What did you do to get the limit removed? 

 

Told NGED (ex WPD) over the phone I had had more panels and a heat pump fitted and wanted the inverter power limit relaxed so it could drive the HP, and the export limit relaxed so I could export any surplus. They agreed subject to a new G99/G100 application but it took a long time to process and much chasing.

 

 

Edited 10 hours ago by sharpener

[sharpener]

5 hours ago, JohnMo said:

. Universal min-SoC targets (work for both summer & winter)
Since WonderWatt will top up if solar looks poor (and supply loads from grid while charging), you just need guardrails so you don’t run flat.

Based on your battery size 13.4 kWh usable:

End of 04:00–07:00 slot: 40% SoC → ~5.4 kWh

End of 13:00–16:00 slot: 40% SoC → ~5.4 kWh

End of 22:00–00:00 slot: 30% SoC → ~4.0 kWh

These are conservative enough to prevent a blackout in winter (since WonderWatt will over-supply if needed because of the higher “baseload” entry), but not so high that you block summer solar.

 

Not sure I understand the rationale of this.

 

Don't the mini-SoC targets mean that the battery is eventually charged to 100% mostly from PV at 15p per unit not exported rather than the grid at 12.65p (at present Cosy rates)? Why is this better? The battery is still cycled from 100% SoC so the depreciation is the same.

[Kelvin]

15 minutes ago, sharpener said:

 

Told NGED (ex WPD) over the phone I had had more panels and a heat pump fitted and wanted the inverter power limit relaxed so it could drive the HP, and the export limit relaxed so I could export any surplus. They agreed subject to a new G99/G100 application but it took a long time to process and much chasing.

 

 

I assume a full G99/G100 application rather than fast track? 
 

I have some options on adding a ground mounted array.  I could add 20 panels on top of a cliff at the bottom of the garden. It has some complication being 80m from where the equipment is and has to cross a burn. Alternatively I have another couple of areas where I could add around 10 panels in two different groups which is beside the garage. Just need to do the sums on all the options. Alternatively I do nothing and just use what I have. 

[JohnMo]

1 hour ago, sharpener said:

 

Not sure I understand the rationale of this.

 

Don't the mini-SoC targets mean that the battery is eventually charged to 100% mostly from PV at 15p per unit not exported rather than the grid at 12.65p (at present Cosy rates)? Why is this better? The battery is still cycled from 100% SoC so the depreciation is the same.

Think it's saying don't bother paying to import then getting hit by DC to AC inefficiencies to get export and effectively get the same real value exporting as importing. The cost difference is so small, just import the minimum you can. If after that if  you have real excess, let that export.

 

On poor solar day and we get plenty, the battery may only partly charge. Today battery has just hit 92% (16:45). Will maybe get a few extra percent or get to 100% and export a little.

 

So far today the battery topped up to 50% early morning, and house took grid supply in during charge period - used 3.2kWh. Have consumed around 10kWh. Generated around 8.5kWh. Have exported 0.01kWh just via stuff slipping by. Just under one full charge today.

 

So the question is how much electricity would I have added to battery to keep at 100%, would the system have drawn in grid during the charge periods? Assume it would still be one charge and PV driving house and exporting?

 

I also like the idea of buying as little energy as I can. But trying different things to see what works. Certainly just charging to 100% is a fit and forget option. 

[JohnMo]

2 hours ago, sharpener said:

 

Not sure I understand the rationale of this.

Been back on ChatGPT, and explored the shite in shite out aspects of AI, asking some basic questions, after a couple of further questions it has spat out a different answer.

Its now saying.Quite like chatting with AI, but you do go rubbish rabbit holes.

• 04:00–07:00: charge to 95% SoC (≈ 12.7 kWh)
  Covers the 07:00→13:00 gap on your coldest days without touching 25.79p.

• 13:00–16:00: charge to 95% SoC (≈ 12.7 kWh)
  Covers the 16:00→22:00 gap (including the 39p peak 16–19).

• 22:00–00:00: charge to 56% SoC (≈ 7.5 kWh)
  Bridges 00:00→04:00 comfortably.

The above basically covers my worse day without solar.

 

So my take on that is after @sharpener valid question, what's the point of charging to 95% you may as well just charge to 100% each slot and forget about it.

 

Have now set the charge automation in GivEnergy Control to work with the Octopus API and automate the charging. Will charge to 100% and then export as its generated.  See how that goes.

[joth]

9 hours ago, JohnMo said:

Asked AI

UK-based Wonder Watt (the smart charging app)

 

The company appears to operate as a small startup based in South Wales, founded by an experienced software developer (Wim) and a solar systems engineer (James) 

deegesolar.co.uk

Interesting. i saw the Deege reference on the website, but AI did a better job of doxing the individuals connected with it. (Yay thanks AI)

 

Deege did my battery install (and forgot to bill me for it for over a year lol), so the name did jump out. They're Kent based.