Why your battery drains before the evening peak even hits

By Declan Osei · October 13, 2026 · 8 min read
Electric vehicle charging cable connected to a car
Photo by Ernest Ojeh on Unsplash

I had a BYD Battery-Box owner message me in July asking why her system was sitting at 4% state of charge by 5:40pm, a full twenty minutes before her retailer's peak tariff window even started. She'd done everything right on paper: sized the battery for her evening load, picked a reputable installer, set up time-of-use export. And yet the one job the battery existed to do, cover the 6pm to 9pm peak, was failing most weeknights. The problem wasn't the hardware. It was a single percentage buried in the app that nobody had touched since install day: the reserve, or minimum state of charge setting.

This is one of the most common battery-plus-solar misconfigurations I come across, and it rarely gets mentioned in the sales pitch because it sounds boring. But it's the difference between a battery that actually does load shifting and one that just quietly burns through your stored solar on the afternoon dishwasher cycle, leaving nothing for the hours that matter.

What "reserve" actually controls

Every hybrid inverter and battery system has a setting, sometimes called minimum SOC, backup reserve, or floor level, that tells the battery the lowest charge it's allowed to discharge to during normal daily operation. Set it to 10% and the battery will happily run itself down to that floor serving household loads, any time of day, as long as there's charge above it.

Here's the catch: most systems treat all discharge the same unless you've also configured time-based control. If your reserve is 10% and you haven't told the system to preserve charge for a specific window, it will discharge for the 2pm load spike (think: oven, dryer, kettle all running during a work-from-home lunch) just as readily as it discharges at 7pm. On a day with middling solar, a bit of cloud cover, a cooler morning that delays the solar ramp, that daytime drain can leave the battery sitting near its floor well before the sun goes down and the real peak begins.

I've written separately about how to set your battery's charge and discharge windows properly, which covers the scheduling side. Reserve is the companion setting, and in my experience it gets ignored more often because installers set a conservative default (often 10-20% for daily cycling, sometimes higher if the customer paid for backup) and never revisit it against the household's actual usage pattern.

The self-consumption trap

Most hybrid inverters, including the Fronius Symo GEN24 and the inverter built into a Tesla Powerwall 3, ship with a self-consumption mode as the default operating logic. The idea is sound in principle: use solar first, charge the battery with the surplus, draw from the battery whenever the house needs power and the sun isn't providing it. For a huge number of households, left completely unmanaged, this works fine.

Where it breaks down is on the exact days you need the battery most. A still, overcast afternoon in late autumn produces a fraction of a clear day's solar. Self-consumption mode doesn't know that 6pm is coming and that your tariff triples then; it just reacts to the current gap between generation and load. If your retainer plan window and your afternoon errands and appliance use are competing for the same stored kWh, the appliance use usually wins by default, because it happens first.

This is a genuinely different problem to the one I covered in why your battery can't do self-consumption and EV charging well at once. That piece is about two loads competing for the same charge rate. This is about the battery's floor setting allowing it to be emptied by the wrong load at the wrong time of day, before the EV or the peak-period load ever gets a look-in.

Reading your own usage pattern before touching any dial

Before I'd recommend changing anyone's reserve setting, I'd want two or three weeks of the app's historical data, not the live dashboard, the daily or hourly export if the manufacturer's portal allows it. You're looking for two numbers: what time the battery typically hits its floor on an average day, and what time your retailer's peak window actually starts.

If those two numbers are close together, your system is working close to as intended, your battery is basically running out right as the expensive hours begin, which is the textbook definition of a well-sized battery doing its job. If the battery is hitting floor two or three hours before peak starts, that's the gap you want to close. And if it's hitting floor by lunchtime, something is more seriously wrong, either the battery is undersized for the home's loads (worth checking against sizing battery capacity for self-consumption vs backup, since these are genuinely different design problems) or the reserve and scheduling settings aren't doing any shaping at all.

Declan's honest take here: installers default to self-consumption mode because it's the lowest-effort configuration to sell and demonstrate on install day, not necessarily because it's the best fit for a customer on a time-of-use tariff. I'd argue most households on a proper peak/off-peak plan are leaving money on the table with the factory default, and almost nobody goes back to check.

Setting a time-based reserve properly

The fix, where the inverter and battery firmware support it (and most reasonably current systems do, including the BYD Battery-Box Premium HVS and the Powerwall 3 via their respective apps), is to stop treating reserve as a single flat number and instead define it per time block. A common approach I use with my own system: set a higher reserve floor, say 40-50%, that applies from midday until the peak window begins, then drop the reserve down to a lower number, 10-15%, for the duration of the peak window itself so the battery is allowed to fully discharge into it.

Outside those two windows, overnight and early morning, the reserve can sit wherever makes sense for your backup preferences. This isn't complicated logic, but it does require going into the advanced or custom schedule section of the app rather than leaving it on the one-size default, and the exact menu path varies a lot between the Fronius Symo GEN24, the Powerwall app, and third-party monitoring platforms layered over a BYD or similar battery.

One thing worth checking before you set this up: whether your plan includes a VPP arrangement. If you're on something like the NSW Virtual Power Plant Incentive, the aggregator may already be dictating discharge events during critical peak periods, and your manually configured reserve schedule needs to sit underneath that without fighting it. Check the aggregator's own guidance before you start layering custom schedules on top, because some VPP software overrides local scheduling during event windows anyway.

Where export limits complicate the picture

None of this happens in isolation from what your inverter is allowed to send back to the grid. If you're on a tight export limit, which I covered in detail in export limits are the real reason your battery doesn't pay off faster, your system may be curtailing solar generation on sunny days rather than letting it all through, which changes how much surplus is actually available to top the battery up in the first place. A reserve schedule assumes there's enough solar surplus to refill the battery to its daytime floor before afternoon loads start eating into it. On a constrained export limit with a smallish array, that assumption can quietly fail on bright days too, not just overcast ones.

It's worth running the numbers for your own system rather than assuming the factory default accounts for your specific tariff structure and export limit combination, because in my experience it usually doesn't; the default is a generic setting tuned for the median customer, not your street.

A few practical adjustments worth trying

Start conservative. Bump your daytime reserve up in 10% increments over a week or two rather than jumping straight to 50%, and watch whether the battery is still reaching a useful state of charge by the time peak starts. If it's consistently topping out at 100% well before midday on sunny days, you've got headroom to push the reserve higher without losing charge to curtailment. If it's barely scraping to 70-80% by early afternoon, a high daytime reserve just means you're buying less grid power at off-peak rates during the loads you'd otherwise have covered from storage, which isn't actually a win.

Check the behaviour across a genuinely average week, not just a run of clear days, since the whole point of this exercise is handling the cloudy, lower-generation days where the gap between battery floor and peak start is widest. The U.S. Department of Energy's guidance on going solar and planning around daily and seasonal generation variability is a reasonable starting reference if you want the underlying logic on why clear-sky assumptions mislead; the Clean Energy Council's consumer guidance on home batteries covers the Australian side of sizing and operating a storage system against a retail tariff.

And if you've got an EV charging overnight or off the same circuit, it's worth revisiting your overnight reserve floor too, since a battery drained flat by 9pm has nothing left to buffer a morning charge session before the car needs to leave. That's a separate scheduling conversation, but it starts from the same habit: actually looking at when your battery runs dry, rather than assuming the install-day settings are still doing their job.

I still check my own system's historical discharge curve every few months, mostly out of habit at this point, because tariffs change, driving patterns change, and a setting that made sense eighteen months ago quietly stops matching how the house actually runs.

— Declan Osei, Battery & EV Charging, IceWarm

Tagged

Common questions

What's a good reserve setting for a home battery used mainly for evening peak shifting?
There's no single number that fits every household, but a common approach is a higher daytime reserve (around 40-50%) that holds charge back until the peak tariff window begins, then a lower reserve (10-15%) during the peak window itself so the battery can discharge fully into the expensive hours. Check your actual usage data before picking numbers.
Why does my battery app default to self-consumption mode instead of a scheduled reserve?
Self-consumption mode is the lowest-effort setting for installers to configure and demonstrate at handover, and it works reasonably well for flat-rate tariffs. It's not necessarily tuned to a time-of-use plan with a defined peak window, which is where a custom reserve schedule usually does better.
Does a higher daytime reserve setting reduce my total solar self-consumption?
It can, if your system doesn't generate enough surplus to refill the battery before afternoon loads start. Watch your charge curve over a couple of weeks of average (not just sunny) days before committing to a high reserve floor.
Will a VPP program override my manual reserve schedule?
Often yes, during declared event windows. Most virtual power plant aggregators reserve the right to dispatch your battery regardless of your local settings during a critical peak event, so check your specific VPP terms before layering a custom schedule on top.
About the author
DO
Declan Osei
Battery & EV charging · Brisbane, AU

Declan writes about home batteries, EV chargers and the emerging vehicle-to-grid category, drawing on his own multi-year EV ownership.

EV owner since 2019; home-charging enthusiast.

More from Declan Osei