Off-grid on the edge: Not quite off-grid, but reliable
Jayne and Cathy Malcholm struggled with power reliability in their location on the edge of Melbourne, so in 2016 they installed battery backup as part of their long-term plan to go off-grid. They describe the system and results.
This article was first published in Issue 143 (Apr-June 2018) of Renew magazine.
We live on the outer edge of Melbourne in a low-energy house which was designed with solar in mind. It has a north-facing roof pitched at 45 degrees to slightly favour winter solar collection. We are on the end of a SWER (single wire, earth return) power line, which over the years has proved to be very unreliable. It was this unreliability that drove us towards a solar + battery system; the power would go off a couple of times a week, resetting our clocks and dropping out the computer, not to mention the inconvenience of sitting in the dark.
We began our project in mid-2015 with some research on solar installers and chose a company that does both domestic and commercial installations. With Jayne having a technical background, she had a lot of questions before we signed on the bottom line. The negotiation took about three months. We kept refining the system specification as we better understood what was being offered and what seemed appropriate for our needs.
If you are thinking of going down the solar or battery storage route, our advice would be to pick a reliable company that is willing to spend the time answering all those difficult questions. It may cost you a little more, but our experience is that you get what you pay for, both in terms of product and installation quality.
Our requirements were based on our long-term goal of being able to go off-grid; we are already off-grid for water. We decided to design for going off-grid while initially keeping our system on the grid to enable us to test whether it would meet our needs.
We wanted the system to be able to power the house for two days without significant solar input and to be able to start a water pump for fire-fighting purposes. As we were intending to install solar hot water collectors, we did not consider PV water heating at this stage.
Choosing the battery type and size
We spent some time looking at which battery chemistry would best meet our needs. Lithium batteries looked impressive with their ability to handle a low state of charge (SOC), but at the time the price was outside our budget. We also had some concerns about lithium batteries in a fire.
We also considered the Aquion salt-based batteries and while these appeared to have a good minimum SOC capability and good environmental credentials, their size was an issue: they are much larger to deliver the same load as the other chemistries we looked at and would not fit the space we had available.
A few other battery chemistries were becoming available at the time, but they hadn’t been on the market long enough to see if they were up for the long haul. In the end, we opted for 32 kWh of lead-acid gel batteries.
Settling on the configuration
Our configuration is a little out of the ordinary in that it uses both AC- and DC-coupled panels. The heart of the system is the Australian-made Selectronic 7.5 kW SP Pro inverter/charger. This device is very configurable with about 400 settings and is definitely not a DIY device, though with Jayne’s electrical background, she has done a bit of ‘tweaking’ around the edges. We’ve found the included ‘SP Link’ software essential for understanding the system operation and monitoring its status.
We added 2.6 kW of AC-coupled panels (with microinverters) to our existing 1 kW system to give us the maximum feed in of 3.6 kW allowed by our supply authority. We sited these panels at the east end of the house, which is subject to early morning shading in winter; we chose microinverters because of the shading issues. We replaced the existing 1 kW system in the centre of the roof with another 1 kW of panels with microinverters, also at the east end of the roof. We didn’t want to use too much of the centre roof space because we were planning to use that for solar hot water collectors in the future.
To help with battery charging when solar radiation is low, we added 3.2 kW of DC-coupled panels towards the middle of the house; we decided we didn’t need optimisers for these panels as this area is only shaded in the late afternoon in the middle of winter. These panels feed into the battery via their own ‘Morning Star’ charge controller.
When the AC-coupled panels aren’t generating the full 3.6 kW of allowable export, the inverter supplements the output by taking power from the DC-coupled panels, but not from the batteries. This configuration provides a higher overall export to the grid, while staying within the export limit. The system is configured to prioritise battery charging over power export, which is not significant until the battery reaches about 92% SOC.
A small addition was our own energy meter that has pulsed digital outputs (and a flashing LED) to indicate whether we are importing or exporting power—something we find hard to determine with the standard smart meter.
The final part of the system is a relatively quiet (54 dBA @ 7 metres) Honda 7 kW (peak) petrol inverter generator with 5.5 kW continuous output. This is automatically controlled by the SP Pro inverter to charge the batteries when they reach a nominated low SOC. The generator does not provide power at peak load times, though this can be configured in the inverter.
We did some modelling based on the expected average solar radiation throughout the year, on a month-by-month basis, together with our electricity consumption (of about 8 to 9 kWh/day) and came up with 6.6 kW of panels as being a good starting point (the ATA solar calculator was not available at that time). Initially we were going to install a 16 kWh battery and add another module later after testing the system and funds became available, but we realised this could have compatibility issues based on battery ageing, so we ‘bit the bullet’ and opted for a 32 kWh battery.
Some might say that a PV capacity of 6.6 kW is a bit excessive considering the limited export we are allowed, but the sizing was based more on charging the batteries during those cloudy winter months than on export considerations.
While this approach has been reasonably successful, we do top up the batteries occasionally via the generator if we estimate that a low SOC will be reached overnight during the generator lockout time, which we’ve set as 8 pm to 8 am.
Having a hybrid configuration, the inverter operates in both on- and off-grid modes depending on the state of the incoming supply. We are connected to the grid, but the inverter is configured to allow export but prevent import, so we are effectively running off-grid. The zero import restriction will be overridden to charge the batteries if the battery SOC reaches a critically low level, but this has never happened.
The penalty for this configuration is the daily supply charge imposed by the electricity supply companies, but the income from the feed-in tariff covers this. For us, this exercise is more of a long-term experiment than about the economics.
In the winter of 2017, we did a ‘real’ off-grid trial. This was accomplished simply by turning off the grid supply at the switchboard. We found that on most days the batteries fully recharged even with moderate solar radiation. On the few occasions when there were more than two consecutive days of low solar radiation, the generator automatically recharged the batteries if they reached the low SOC set point.
We have an all-electric house and were initially expecting that our solar/battery system would meet our general household power requirements, with our water heating requirements continuing to be met by the grid overnight. We operated this way for a few months while we thought about alternatives.
We could no longer use our reverse-cycle air conditioner for room heating because of the power required from the battery, so we decided to put in a wood heater with a wet back for water heating. This was an economical approach as we can source the wood from our property, and it solved the issue of water heating as well. We found that if we ran the fire all day, we could heat our water with ease, so we stopped boosting the hot water tank at night from the grid.
When the weather started to warm up, this approach was impractical as we didn’t need to use our wood heater for room heating as much, if at all. As our water tank has main and boost elements, we replaced the 3 kW boost element with a 1 kW element which we manually switch on during the day if the solar radiation is sufficient. While it only heats water at the top of the tank we have found that it reaches the temperature required and meets our fairly small hot water needs.
A 1 kW element may seem small, but we chose it because there was a relatively small amount of water to heat and we felt that our system could cope with a 1 kW load on most days. The thermostat usually stops it heating by the end of the day, but we still turn the boost element off in the evening to avoid it coming on the following day if there’s insufficient solar radiation. We have minimised the tank cooling by adding additional insulation around the tank. In practice this has worked out quite well. We’ve also developed an alternative plan to automate the switching based on battery SOC and energy available, using a custom controller that monitors currents in different parts of the system and makes logical decisions as to when to switch the water heater on or off.
We have also looked at another alternative, using a power diverter to heat the hot water. Instead of connecting the sensor to the incoming grid circuit (which works only for on-grid mode), we would monitor the current from the AC-coupled panels in conjunction with the battery SOC. The SP Pro inverter can be programmed to operate a relay contact at any nominated SOC, so we could connect this relay output to the power diverter current sensor and have it only operate when the SOC reaches about 92%. This would allow the water to be heated only once the batteries reach their required state of charge and there is sufficient solar energy being generated.
So how has the system performed? Since it was commissioned in February 2016, we have only taken power from the grid for off-peak water heating. Since the wood fire was installed in September 2016, we’ve taken no power from the grid and our export has more than covered the daily supply charge.
One negative we’ve found with the current configuration is that if we run without the grid connected (i.e. no export possible) and the batteries are already at a reasonable state of charge, we generate too much power. In this case, the inverter shuts down the microinverters by raising the supply frequency to 55 Hz. This is not a major issue, except all our clocks/fans run fast. Our planned custom hot water controller or a power diverter to use the excess power will help reduce this occurring. Generally, we run quite well in off-grid mode, with the AC panels able to power the house and charge the batteries as they’re on the ‘house’ rather than the ‘grid’ side of the SP Pro inverter.
We haven’t monitored the amount of fuel used by the generator, but it has run for about 50 hours since February 2016. This includes regular automatic maintenance operation.
As the inverter is not marketed as a UPS (uninterruptible power supply), there can be slight dropouts when the grid supply fails. Sometimes the changeover is seamless and we believe it depends on the nature of the grid failure.
Would we do it again? Yes: it has been an interesting learning curve for us. When we work through the hot water issues, we would like to go completely off-grid.
In many ways, this has turned out to be a very expensive exercise that will never pay for itself while we are here, but it is nice to have an electricity supply that does not leave us sitting in the dark.