I gave a presentation at the largest energy conference in Australia, All Energy. In my presentation I talked about the future of home energy with AI and available data.
Sam Bendat
Originally Published: Nov 05, 2024
Updated: Nov 18, 2024
The presentation itself mainly focused on AI as a tool to create value for homes based on their consumption history. With or without AI, there is a lot of value to be created by analysing consumption patterns and helping people make informed decisions based on that data.
In the last article, I talked about how a conversational AI like ChatGPT could be used to create a quick energy assessment of my home for free.
A second example is using our data to design a solar system that considers our consumption patterns, roof size, solar energy hitting our roof and a few other variables.
In the above slide on the left we have a dozen or so forms of data feeding into an AI model that will create a solar system for a roof. All of the data can be obtained in a few seconds without installing devices or having someone visit your home. Naturally, each homeowner has to consent to share this information, though.
The value here is that the system is automatically designed with best practices in mind using the same data every industry software product uses. It can then estimate the dollar savings the system will achieve over time. The output on the right is a system designed on the roof according to the consumption patterns of the home, maximising the value and return on investment for the system.
There are some products out there that can do this for installers, but it would be great to put this information in the hands of homeowners so they can make informed decisions around solar—creating a better understanding of how a solar system will benefit the homeowner, what size solar system and what kind of physical setup on the roof we can expect.
We can also do a similar analysis for batteries in the home. Long-time readers will know I have had a go at battery analysis on multiple occasions. To get a rough idea of how a battery might perform in a home, we can feed an AI model a bunch of new data sets to figure out how the battery might perform in the home.
If we consider homes that already have solar we'll need to access the production and consumption of that system. From there, we can see if further optimisations can be made around charging and discharging the battery at different times, to optimise its value for the home.
Like for example, if the energy plan for a home has a cheap off-peak rate from noon to 4pm, it might make more sense to buy that energy from the grid and store the solar energy in a battery to use later that day. Maximising the return for the home.
To this day, I still don't understand how anyone can rely on an energy bill to create a personalised quote for anything energy-related. Bills simply don't have enough information on them to keep us up to date about what is going on in our homes.
Imagine you just installed solar and a new hot water heat pump, and now you want to find an energy plan to lock in a rate for the next year. Of course, last month's energy bill isn't going to help. What we need to do is forecast the best plan for your home given the upgrades to the home.
Predictive modelling, in this case, can be based on analysing the consumption patterns of 1000's homes. Once we understand how the consumption patterns of a home changes, we can give a far more accurate recommendation on the best plan available to each home. Its a forward looking analysis based on real consumption history and pattern recognition rather than a shot in the dark with energy bills.
At any one point, there are around 13,000 energy plans available in the national electricity market. Its an overwhelming task to find the best plan for their home.
The final example is handing over the management and control of the energy-hungry devices of the home to create an autonomous system that manages our consumption for us.
This can also be done without needing to install another device that costs extra to place in the middle of an existing home energy system. It will be possible to be done remotely and in just a few clicks without even getting up from the sofa. Once connected I'll be able to set some basic rules on how I want my home to be run then all of my devices will automatically run themselves while optimising their consumption according to my energy rates, the weather, the rules I set, and more.
With the above solution data can be used from the smart meter in the home, the solar system, the EV in the garage, the split system, the hot water heater and more, remotely accessed over the web. The homeowner is in control but to keep costs low and the comfort of the home high the complex calculations are handed over to an app to make it all possible.
Given how home energy is moving where devices are accessible, readable and controllable from an app, this kind of full system integration is really a matter of when instead of if. Exciting times ahead.