Geo-specific weather data and artificial intelligence from IBM and The Weather Company are helping solar inverter manufacturer Selectronic efficiently store energy in solar batteries, increasing the value of the expensive renewable energy storage devices.
Since the early 80s, Selectronic has been designing and manufacturing solar inverters. In this case, the inverters, the brains inside the actual energy storage battery system, control the renewable energy that flows to and from solar batteries.
Similar to the batteries in mobile phones, solar batteries deteriorate over time and fail to hold a charge. In a blog post, Dr Julian De Hoog, Technical Lead – Energy & IoT at IBM Research Australia said solar batteries are still relatively expensive so it is essential to operate them in the best possible way to ensure they provide an appropriate economic return.
For the past five years, IBM Research has been working with Selectronic to learn how batteries, solar panels, inverters and cloud-based services can improve these systems.
De Hoog said his research team did a preliminary study to see if it would make economic sense for homeowners to install large batteries in their homes.
“We thought it would be a fairly straightforward thing to analyse, but fairly quickly we realised it’s a much more difficult problem to solve than you think,” De Hoog told Which-50.
“The amount of value that a battery gives you depends entirely on how you run it. And how you run it means you can use a battery for multiple different purposes and you can use it to store your excess solar energy.”
The team needed precise weather data and the help of AI to accurately predict power generation and optimise the battery storage system.
Rod Scott, CEO of Selectronic said, “One of the biggest consumable parts of an energy system is the battery, in terms of the actual longevity of the battery but also in terms of the efficiency of the battery. Being able to improve those two things makes a big difference to any system.”
Selectronic says AI and weather data from IBM Research can help extend the life and efficiency of lithium batteries used for solar storage.
How it works
The Weather Company’s data is precise enough that it can give unique forecasts for individual neighbourhoods and buildings.
The artificial intelligence in the system uses this weather data to operate the inverters in an optimal way to ensure they provide an appropriate economic return and store the most amount of solar energy for that particular day.
De Hoog said this detailed data is valuable because the way users get value out of battery changes from one house to another. As the weather varies for each area, solar panels need to be optimised for each location they are in.
To get the most out of the solar panels they need to forecast 24 hours ahead so they know how to optimise the battery usage for the following day, De Hoog said.
“When we generate those forecasts they have to be unique to specific houses and businesses, both in terms of how much solar they’ll be generating and in terms of how they’re going to be using energy throughout the day,” he explained.
The Weather Company forecasts include standard weather variables such as temperature, humidity and wind speed pressure. And most importantly they also get solar irradiance data, which is key in determining how much energy the panels can take during the day.
The weather data gives the user minute by minute insights into how much solar is being produced, according to De Hoog.
“When you look at that you can actually see at a very high resolution the impact of clouds, trees and roof features. You can see these time series of solar generation and they’re very different from one house to another.”
He said how much solar can be generated depends very much on whether the solar panels are facing east or west and whether there are neighbouring houses that cast a shadow on your panels.
“A standard generic forecasting approach doesn’t work, you really have to look at every single house and sort of learn what the pattern is and then take into account weather and the historical data to create a forecast that’s really specific to that house.”