Picture this. You're an engineer, and you have unreliable electricity. You've experienced multiple long power outages in recent years (2022, 2023), and you know it's only going to get worse in the future, with increasing amounts of extreme weather and worsening energy scarcity. You also have some doubts about the grid coping with the influx of EV chargers. What do you do?!
The logical solution to this problem is to have your own means of generating electricity. For me, that has to be primarily renewable. Unless you happen to live beside a river that never stops flowing and have the civil engineering skills to build a hydro dam (I have neither of those things), then you're probably looking at wind or solar power. But the sun doesn't always shine, and the wind doesn't always blow, so you need energy storage.
Energy storage also helps if you don't have vast renewable energy sources that can cater to every occasion, and need to use fossil fuels as a backup. Energy storage can allow you to efficiently use a petrol or diesel generator to charge a battery, and then use power from the battery instead of the generator. The key to this is that generators have a considerable overhead just from idling, so using a generator to quickly charge a battery and then shutting it off is much more efficient than running your loads directly from the generator all the time, assuming your loads are small compared to the generator's size. Doing this can therefore reduce fuel consumption, and emissions, and of course noise.
Energy storage is also useful even if you don't have solar panels or a generator at all - just storing energy from the grid for use in an outage can be a useful endeavour.
There are really two or three options that people would normally consider, to get a means of electrical energy storage:
Let's look at these options in more detail.
Option 1 is great if you don't want to deal with many technical details. You buy an off-the-shelf system, and have it installed by a licensed professional. The device comes with a warranty, and the installation is signed off under law, so you are covered by law for just about any eventuality until the warranty expires. But it costs a lot. A LOT. Prices are coming down but as of this writing, a Tesla Powerwall costs around $17,000 excluding installation, or approximately $20,000 all-up (in NZ). For this, you get 13.5kWh of energy storage, with a continuous output power capability of 5kW and peak of 7kW. There are many other options on the market and the prices vary, but this is the ballpark for these systems.
Hardwired systems can generally be configured to work with existing or new home solar systems, and configured to back up certain circuits of your home. If your installer configures it correctly, you will be able to charge the battery from solar power while the grid is off - because your solar system is connected to one of the battery backed up circuits and can feed power back to the battery.
However, you can't change any of that on the fly. It's hardwired, so you can't modify the setup yourself, and the more customisation you want for the installation, the more you'll pay, so expect that $20k price tag to increase by a few grand if you want an unusual installation. You also need permission from your local lines company and the system needs to comply with a wide range of regulations that restrict where you can install it and how you can operate it.
These systems also are fundamentally cloud-connected. Essentially, that means they require an internet connection if you want to control them, so if you lose internet connection they will stay in the last operating mode that you set. This could be very tricky during an extended power outage, if you also lost your internet connection. There have also been reports that Tesla and other companies will void the device warranty if there is a sustained loss of internet connectivity and Tesla confirms the potential for this here. Thus sucks big time. I am a big fan of 'right to repair' and 'right to ownership'. This isn't it. If there's one lesson we should have learnt from the extreme weather events of 2023 (so far!), it's that you cannot rely on having an internet connection during a big power outage. Cell towers have limited battery backup, and there were numerous reports of cell tower backup generators being stolen. Many people lost internet connectivity for almost as long as the power was off for. If you cannot control your own system during such a situation, arguably you don't own it. To my knowledge, Powerwalls do have a means of local control, but it requires technical knowledge and is not easy to use by default. And of course, when the device is online, it will constantly receive software updates from the manufacturer. This means that at the manufacturer could easily disable features in the future, such as local control. To some, this statement may sound hyperbolic, but this stuff actually happens. Manufacturers have 'life cycle plans' for the products they sell. At the end of these 'life cycles', they frequently make decisions to terminate product support in ungraceful ways, sometimes even rendering devices unusable (I'm looking at you Amazon, and again here). For anything cloud-connected and receiving software updates or relying on external services, I can't shake the feeling that you don't really own the system. And I fundamentally dislike the idea of connecting an energy storage device to the internet. The only reason it truly needs to be connected to the internet is because it needs software updates, which it only truly needs for security reasons, because it's connected to the internet and could be hacked. An offline system would lack some convenience features like remote control, but would be secure and reliable because it couldn't be hacked and nothing could get broken in future software updates.
Not to mention that if such a device fails outside of warranty, you'll have few options for repair. If you have a component failure in a hardwired home battery system, it's quite unlikely that you'll be able to get a technician to perform a component-level repair on such a device. If you're lucky, they'll be able to replace an entire module, such as an inverter board. If you're even more lucky, the cost of doing so might be feasible. But that's far from guaranteed, and often times a single component failure in a device like a home battery system (but this also applies to heat pumps, EV chargers, etc.) will render the device unrepairable and it will have to be replaced. If that's not covered under warranty, you are completely screwed. If the battery cells are integrated into the device you may end up having to replace them too, just because of one component fault. That's not only expensive, but a terrible waste. Is this just the way things must be? I think not. Manufacturers do not make it easy to do component level repair, out of both laziness and conspiracy to increase sales. They will not provide spare parts, schematics, or resources that anyone, including a professional technician, could use to do component level repairs. The focus is always on module or device replacement, which is an approach that hurts the consumer and creates waste. That's not 'right to repair', and I don't like it.
Option 2 is great if you're a bit more tech-savvy and want more power independence. You don't need to hardwire the device, so you have more flexibility over where you can use it and how you can configure it, and no installation costs. The EcoFlow Delta Pro is one of the leading contenders. It provides 3.6kWh of energy storage, with a continuous output power capability of 3.6kW and peak of 4.5kW. This is enough for many intents and purposes. However it costs over $7000 in NZ at the time of writing. Comparing the cost to that of the PowerWall, you are paying a lot more per kWh of energy storage (approximately $1944 vs $1481 per kWh), even without installation costs.
There's also a problem if you want to set up your own renewable power system with such a device, or use a generator. Although the EcoFlow Delta Pro and similar units have the capability to connect to solar arrays, they do this through a dedicated DC connection, and don't state support for AC-coupled solar charging. This means you can't use them with a grid-tied solar system like you can do with a Tesla Powerwall. The fundamental limitation is that these devices appear to be unidirectional inverters, meaning you shouldn't attach grid-tie microinverters (e.g. solar panels) to their outputs and expect the batteries to charge.
If you have a grid-tied solar installation without hardwired batteries at home, that system must shut down during a power outage. This is for two reasons. The first is to prevent power flowing back to the grid, but that can be solved with transfer switches and the like. The second reason is because there is no storage to stabilise your system and keep the microinverters from shutting down. If you solved the first issue, say with a transfer switch and inlet similar to what you'd use to connect a generator to power the whole installation, then theoretically connecting a bidirectional inverter to that inlet would let you run your own off-grid setup, without a hardwired battery. However, these portable systems cannot perform that role - they can act as a grid-forming inverter to activate your grid-tied solar systems, but likely cannot accept reverse power flow, and will shut down rather than charging the batteries. This sucks.
Also, these devices typically use the same converter for charging and inverter operation (which makes you wonder... is it bidirectional after all? But they don't advertise compatibility with grid tied solar). This is fine if you are planning to charge from the grid, and discharge to a load, at separate times. But if you want to power critical loads during an outage, and simultaneously recharge your battery at the same time from a generator without dropping the loads, then you are out of luck. Because of the single converter architecture, these devices don't allow you to isolate the grid (or generator) from the load while charging the battery, and there will be a transfer time of possibly 30 milliseconds when switching in between with a phase discontinuity. A lot of portable generators also put out a highly distorted waveform under load. You'll have to run your critical loads directly off of this while charging, if you use a portable battery of this type. This sucks too.
And the same cloud-connected problems apply as for the hardwired battery systems, generally. The EcoFlow unit does allow standalone communications with a phone at least, which is good. But what if the app goes unsupported, or gets removed from the app store for whatever reason? These app-enabled systems seem smart and reliable in the moment, but in the long term they are planned obsolescence. I don't like it.
And the same goes for device repair: a single component failure could render the device unfixable, or make repair prohibitively expensive. The warranties on portable systems are also typically shorter than hardwired systems.
Option 3 is where it gets interesting, if you're technically savvy enough. This is more similar to option 2 than option 1, because it won't be hardwired, so you'll have maximum flexibility. But it also may cost less than option 2, if you don't factor in your own time in setting it up, because you can source your own battery cells at the lowest price. If you browse websites like https://diysolarforum.com/ you will see how it is possible to import LiFePO4 cells from a variety of sources in China to most places in the world, for substantially less than what you pay per kWh with commercial systems. However, then you've got to add in the cost and setup of an inverter and charging system.
The first problem is getting parts that will work with each other. Bidirectional inverters exist, but often are for industrial applications and may be difficult to source and use, not to mention expensive. Think $4000 NZD or thereabouts for a good bidirectional inverter the size of the EcoFlow unit, such as the Schneider Electric SW 4048. I'm sure there will be lower-end brands that are slightly cheaper, but you get the idea.
Your inverter also limits what your system is compatible with. Different solar installations, with different microinverters, will behave differently. Some Enphase microinverters require proprietary signalling, while some will respond well to frequency shifting of the AC power to control the power flow. In this regard, some will have a frequency threshold over which they stop operating, while some will linearly reduce their output power with an increase in frequency, allowing precise control by the grid-forming inverter. Your grid forming inverter needs to play well with the microinverters to allow the whole thing to work.
And have fun finding out exactly what your particular inverter or microinverters will do, unless you can find out exactly what standards they comply with and have access to those standards (that costs money too). The consequence of this is that you may well find that after spending $4000, you have an expensive inverter that won't work with your microinverters, and have to replace one or the other to make the system work. The configuration of most microinverters will be locked by the installer for grid compliance too, which will prevent adjustments. You know what I think by now - if you can't change the settings, you don't own it. So, I've come to the conclusion that nobody really owns their own grid-tie solar systems. And that sucks.
Most inverters also tie you to a specific battery voltage, like 48V for the aforementioned Schneider Electric SW 4048. This starts to dictate your choice of battery cells and the capacity of your battery bank. If I used, for example, 100Ah cells, I could get 4.8kWh. What if I want 6kWh? There are limitations to the battery cells available, and the next option up is probably three strings of 50Ah in parallel, for 150Ah, and that's 7.2kWh, but also possibly an impractical number of cells. If you wanted to stick to large cells, then the next common option is probably over 200Ah, or two sets of 100Ah in parallel. That's way more than 6kWh! The common battery sizes available define what capacities you can choose, so when your inverter also dictates the total battery voltage, you are left with limited customisation options. 12V, 24V and 48V inverters are common, but above that voltage it's difficult to find anything outside of industrial products that are hard to get hold of, let alone find documentation or support for.
This is actually the option I'm going with. Here is why.
Picture being the engineer, again, and look at options one to three: what do you do? I guess you'll also have to remember that I'm a student, so I don't have that much money, and I don't own the property I live in. That certainly rules out a hardwired system, and anything that isn't good value for money, like most of the standalone portable battery systems.
Spending thousands of dollars on an inverter isn't my idea of money well spent either. It's hard to guarantee the thing would even work well, there are a bunch of restrictions like input voltage and output controls, and if I break it, its proprietary nature may make it unfixable.
So I'm building my own inverter! While the costs will end up similar to buying one, they can be partly justified as an educational expensive (personal, not University related, of course). And this way, I can design for the specific input voltage I want, and I get full control over the way it handles microinverter control so I can make it work with the particular microinverters I have access to (or different ones in the future).
I think the most important difference is that I will own it: I will have the means to fix it at component level if it breaks, and the means to customise its behaviour to work properly in the way I need it to. It also won't be connected to the internet, so don't bother trying to hack it. And it won't have planned obsolescence built in.
This really only makes sense because I'm in this exact field trying to learn about these systems - but it does make sense, for me, and I think it has some obvious advantages over the other options.
I acquired 8kWh of LiFePO4 storage for approximately NZD $4000, by importing 24 EVE LF105 cells from China. That price may seem somewhat expensive to many, but the costs of importing to NZ are substantial, especially when you import less than a pallet load of goods, like I did. A decent part of the price was simply getting the goods through the local port and through customs. It's still far less than buying similar cells directly here in NZ, and learning how to import goods was also important to me.
And even though I ended up spending about another $4000 building my inverter, it works out to about $1000 per kWh - still less than buying any commercially available system. Beating the cost wasn't my goal, but it certainly sweetens victory.
My inverter has a nominal input voltage of 72V and can also act as a grid connected charger, and as a UPS system. More details are available in the next post.
None yet!
Display Name
Email (optional, not displayed)
Website (optional, displayed, include "https://")
Comment (up to 1000 characters)