A recent article in the Salzburger Nachrichten tells the story of Klaus Strasser, an electrical engineer who built a house in St. Gilgen with no connection to the public grid at all. His own PV and battery system covers everything: heating (via a ground source heat pump), hot water, and household loads for a home that also serves as his office. It is a genuinely impressive piece of engineering. He put about 100 m² of panels on the roof and the facade (21 kWp in total), designed a custom two-battery setup to get clean 400 V three-phase power out of two normally incompatible battery chemistries, and sized his battery bank to ride out seven to ten days without sun. Eight years in, the system has apparently failed him only once, for 90 minutes.
Reading that made me want to check the numbers for a more ordinary case: a newer, well-insulated but not passive-house-standard home somewhere around Klagenfurt, heated with a heat pump, with an EV added a bit later. Would going off-grid make sense there too, or does Strasser's approach only pay off because he built a genuinely oversized, purpose-designed system from day one? Let's do the math.
The setup
Assume a single-family house built around 2017, decent insulation, heat pump for both space heating and hot water. Annual electricity consumption is 6,500 kWh, mostly concentrated in the colder half of the year because of the heating load. An EV gets added later, bringing annual consumption up to about 8,500 kWh.
The roof comfortably fits a 10 kWp system (part of it is shaded by a tree, but let's not worry about that for now). Using a real Meteonorm-calibrated solar yield simulation for Klagenfurt, that system would produce about 10,953 kWh per year, which is more than enough to cover the 8,500 kWh of annual consumption on paper. So far, so good.
The catch is that annual totals hide the timing problem. Here is the average daily production and consumption per month:
In December and January, this 10 kWp system produces only about a third of what the house needs on an average day. From March through October, it produces a comfortable surplus, peaking at nearly two and a half times the daily consumption in June.
Two different reasons for the winter deficit
It's worth pulling apart what actually causes that winter dip, because it's really two separate effects layered on top of each other.
The first is simply shorter days. Less daylight reaching the panels each day, a direct consequence of the Earth's axial tilt, and it happens everywhere that has a real winter. This alone accounts for the dip in the production curve above.
The second is heating demand. Because this is a heat pump house, the electrical consumption side rises in exactly the same months that production is at its weakest, which is the worst possible timing for a system that has to balance daily.
These two effects are related, but maybe not in the way it first looks. It's the reduced solar input of shorter, lower-angle winter days that is the underlying reason it gets cold in the first place, in terms of the surface energy balance. The heating demand is a downstream consequence of that same seasonal solar minimum, and how severe it gets depends on latitude: further from the equator, that minimum is colder.
Which has a nice implication: the heating penalty isn't a fixed law of going off-grid, it's a property of a specific climate and latitude. Move the same house far enough south, or let a few more decades of global warming run their course, and the picture eventually flips: no more heating load in winter, but a cooling load in summer instead, right when PV output is at its peak. At that point the seasonal mismatch would actually shrink rather than grow, since summer cooling demand and summer PV surplus line up nicely, unlike winter heating demand and winter PV deficit. The math in this post is specific to our climate today, not an inherent limitation of PV autonomy as such.
What it would take to close the winter gap with PV alone
If the house is grid-connected, this seasonal mismatch barely matters: the grid absorbs the summer surplus and covers the winter deficit, and net metering or feed-in tariffs settle the difference. But full autonomy, by definition, is not allowed to rely on the grid. So the system has to produce enough on an average December day and an average January day to meet consumption on those days too, not just meet the annual total.
Scaling up (keeping the same roof mix and orientation) to hit that daily balance in the two worst months gives:
| Month | Consumption/day | Production/day at 10 kWp | Required system size |
|---|---|---|---|
| January | 26.7 kWh | 12.1 kWh | ≈ 22 kWp |
| December | 26.7 kWh | 10.1 kWh | ≈ 26.5 kWp |
So the array would need to be about 2.6 times larger than the roof-only 10 kWp system, roughly 145 m² of module area (around 58 panels at 460 Wp each). That is already well beyond the 121 m² of usable roof area assumed here, which is exactly the kind of constraint that pushed Strasser toward using vertical facade surfaces in addition to his roof. In our hypothetical case, closing the winter gap the same way would mean covering most of the south, west, and east facades with panels as well, on top of the roof.
And even if enough surface area were available, a 26.5 kWp system built to satisfy December would massively overshoot the rest of the year. At that size, an average June day would yield somewhere around 130 kWh, against a consumption of about 20 kWh. That is roughly 6.5 times more than the house can use on that day. Without a grid connection to sell that surplus into, most of it would simply have to be curtailed, since there is no affordable seasonal storage technology that lets a household charge up in summer and draw it down five months later in winter. Multi-day battery buffering, like the 7 to 10 days Strasser built in, helps with cloudy stretches, but it does nothing for a structural seasonal shortfall that lasts weeks.
So what would actually help?
To be clear, none of this is a knock on what Strasser built. Going fully off-grid is a legitimate and interesting engineering goal, and if you are willing to over-build the PV array, use every available surface including facades, and invest in a serious multi-day battery, it clearly works, his own house is the proof. It is also worth remembering that he never fully gave up the option either: a grid connection was prepared during construction, just never activated.
For a house that does not want to go that far, the more sensible response to this same winter problem is usually not more PV area, but one of two other levers:
- Stay grid-connected and let the winter deficit be covered by grid draw while the summer surplus gets fed in. Given that the "right-sized" 10 kWp array already overproduces by a factor of about 1.3 annually, there is a reasonable economic case for feeding that surplus back into the grid rather than throwing it away, since the cost of a modest grid connection is easily justified by the value of the exported summer energy.
- Diversify the heat source for the winter months, for example with a wood or pellet backup, so that the electrical load in December and January drops closer to the summer baseline. This attacks the actual root cause (a heating-driven winter consumption peak) instead of trying to out-build it with panels.
Winter autarky bought purely through PV oversizing is, in most residential cases, the economically weakest of these options. It is a fun problem to size on paper though, and seeing how quickly the required array grows once you demand daily balance in December is a good reminder of just how seasonal solar really is at this latitude.
