Battery Power-Sharing

By Nithya Menon
May 17, 2021
The Pod now enables transmission of underutilized battery power to wherever it’s most needed at night.

It is with great pride that we celebrate one of our most eagerly anticipated features of all time: battery power-sharing (we’ll refer to it as BPS from now on). BPS builds on our existing smart power-sharing algorithms, and instead of only redistributing excess solar power during the day, the Pod now enables transmission of underutilized battery power to wherever it’s most needed throughout the night – usually towards households that are running energy-intensive productive appliances. A lot of features came together to make BPS possible, and now we’re excited to see families using more high-power productive appliances all while experiencing fewer blackouts and costing the grid owners less on equipment.

Sounds too good to be true? Let’s dig in.

Flow of power when battery power is being shared between households

Why is BPS a game-changer?

And why was it one of the most sought after features that we’ve been working on? Because BPS plays a key role in achieving reliable 24/7 power for productive appliances without massively oversizing the systems, resulting in saved costs, higher revenues, and increased grid uptime.

Households using productive appliances experience the greatest impacts of electrification while generating the most revenue for the grid developer. But in order for productive appliances to run through the night, the local battery capacity was a limiting factor even when there was available capacity elsewhere in the network. Conceptually, BPS made sense as a way to access that inaccessible, wasted capacity throughout the night rather than always needing bigger batteries at every productive household. We built simulations to test various grid configurations, load profiles, and BPS limits/thresholds and saw that BPS indeed reduced blackouts significantly, especially as energy demand grows.

For example, take three homes, each with a standard 100Ah battery, and then one house increases its daily usage to 2.4kwH, a second house increases to 1.8kWh, and a third home maintains a 600Wh load. BPS can reduce blackout hours by 88% and bring the average time spent in blackout from 15% down to 2%. The exact reduction of blackout hours will vary depending on the specific grid conditions, but clearly, the effect can be substantial. One caveat is that BPS will have less impact if all systems are oversized or if all homes grow their load usage simultaneously. But in reality, homes scale up at different times and developers want to limit oversizing, so BPS can make scaling networks smoother and more profitable.

Considering how important it is to facilitate scaling networks in order to achieve profitable and sustainable grids, we also modelled how other product features could help reduce the required additional CapEx over time, to sanity check how impactful BPS could be.

Savings as % of Total Capex Cost over 5 Years

Using a mixture of load profiles typically observed in off-grid communities and 20% of households increasing their load profile annually over 5 years, we see that sharing excess panel power does reduce CapEx costs by redistributing excess power, but doesn’t have nearly as big an impact on upfront costs as BPS. With BPS, the network can operate on smaller batteries, dramatically decrease wasted power, and facilitate easier load growth, thus reducing CapEx costs by as much as 16% (modelling assumed 40% households using 200Wh/day; 40% households using 400Wh/day; 10% households using 800Wh/day; 5% households using 1500Wh/day, and 5% households using 2000wh/day).

Finally, in addition to increasing grid uptime and reducing required CapEx, BPS has one more long term benefit. On top of being susceptible to outages, productive systems face reduced lifetimes, as the batteries in these homes get thrashed from frequently being drained to the point of blackout. Lithium batteries are more durable to greater depths of discharge compared to lead-acid batteries, but the lifetime charge cycles will still be reduced if the battery is fully discharged often. In this example, we model how a single house’s battery charge is affected based on BPS being available.

House 1 Battery SoC with and without Battery Sharing

Similar to previous models, we assume this house has increased its load consumption via a productive appliance and is grid-connected to at least one other home maintaining an average load profile. We again see how without BPS, the battery of an individual system cannot support the productive appliance and would trend towards blackout without being upsized, and we also see the degrading effect on the battery’s state of charge. BPS maintains healthier states of charge across the grid thus prolonging battery lifetimes and lowering battery replacement costs for the grid owner.

Our models continuously validated BPS as a key feature to reduce costs and increase grid reliability, especially as networks scale over time. But modelling aside, the implementation was no simple task. Mastering the ability to distribute power between our pods was one of the biggest technical battles we faced in the early days (check out a past blog with plenty of juicy details about the electronics and control algorithms). And even after those successes, we faced constraints that made progressing to BPS substantially more challenging.

Up until recently, lead acid batteries were the only affordable and easily accessible batteries, but pulling power out and pushing it back into a lead acid battery (known as the round-trip efficiency) is only 70-80% efficient. As a result, it’s highly inefficient to pull power out of one battery at night to send over the grid cable and into another house’s battery. The combined losses make it more wasteful than beneficial. By initially restricting the algorithm to only send solar directly to another battery over the grid, we cut out multiple in/out battery transitions, but this also meant restricting the hours per day the algorithms could be active.

Price of Lithium-Ion Battery Packs per kWh

In the past decade, lithium prices dropped, the larger upfront investment started to have a practical payout, and the > 95% round trip efficiency along with greater available capacity made redistributing power at night more impactful. So in 2020, we redesigned our pods to charge lithium batteries. Lithium charging not only required a whole new algorithm but also needed careful cooperation between our controls and the inbuilt battery management systems (BMS) of the batteries. By the end of the year, partners in Cambodia rolled out the first Okra grids using entirely lithium batteries, getting us one step closer to BPS.

After gaining confidence and stability in our lithium algorithms, the next barrier was a reliable state of charge calculation. The daytime sharing algorithms rely on the battery’s charging phase to make decisions, but at night there would be much less info to feed the algorithm. To fill this gap, we built a Kalman algorithm (see blog here) to give the BPS algorithm access to battery state of charge data, ultimately useful to define key send/receive thresholds in the logic, so pods know when they have excess power to share or when they are on track for a blackout and require support. At the lowest level, panel sharing and BPS share a common foundation in the code and use the exact same hardware components, but the higher level logic of when to share, when to receive, and how much power to push is where innovations in SoC monitoring played a key role in making BPS possible.

BPS has been deployed for about a month now, and preliminary data is already promising. Across all our lithium projects, we now see 40% homes contributing power at night and 25% of homes receiving, resulting in a 19% increase in usable power redistributed via the grid to support homes running low on power at night.

Battery Sharing vs Panel Sharing Distribution

Distribution of Homes Sending / Receiving Battery Power

As we collect more data, we’ll not only expect improved network uptimes but also an increase in load consumption now that high-load systems are performing better. When it comes to tackling challenges in off-grid electrification, no single product feature can solve everything, but BPS is a huge win – enabling energy companies to finance more productive loads and deploy more profitable minigrids, and ensuring households not only have access to reliable power but also access to the important and life-changing opportunities electricity can unlock.

Thank you for coming along on this journey and helping us celebrate this milestone! As always, don’t hesitate to reach out with questions, comments, and ideas!

Use Okra

Learn more

Nithya Menon is an engineering graduate from Harvey Mudd College that has spent her career developing technology targeted towards empowering marginalized and developing communities worldwide. She has been pivotal in designing Okra's key power-sharing algorithms, IoT firmware, and grid management software - and now drives the direction and strategy of Okra's technology as Product Development Lead.