Using solar monitoring to detect bushfires highlights value of data

Bushfire smoke on a sunny day. Researchers have developed a way to detect bushfires using solar monitoring data.

Researchers at UNSW have come up with a way to detect bushfires using solar monitoring data. They have built a model to estimate the amount of fine particulate matter (PM2.5) in the air by comparing actual solar generation with the expected solar generation on a clear day (based on weather and irradiance data). This basically turns solar panels into (low fidelity) air quality sensors. 

What makes this interesting is that there are a lot more PV systems in Australia than weather stations with air quality sensors. Even though the accuracy of the PM2.5 estimates from solar data is lower than that measured with air quality sensors, the much more granularly distributed nature of PV systems would make this at least a great complementary system to existing bushfire detection methods.

This study was conducted with data from 160 residential PV systems in NSW during the 2019–2020 Black Summer bushfire season. Turning this study into an actual bushfire detection and monitoring system requires (near) real-time data from PV systems spread across the country.

Energy Data as a Service

This is a great example of what happens when high quality energy data is available. People find ways to use the data  beyond what the curators of the dataset would ever have thought of. 

Yet, the availability of energy data has historically been a challenge. Acquiring residential electricity consumption data is expensive; it involves installing monitoring hardware and ingesting, storing and managing the data. It’s something that every energy research project needs, but something no individual research project can afford to do at scale.

In the Australian context the Smart Grid Smart City dataset was once the main data source on residential electricity use. It consists of 3 years of historical data from (just) 300 households. While it’s great that this was (and is) an open data source, its relatively small size and historic nature limits its usefulness.

‘Energy Data as a Service’—a concept Wattwatchers team members have talked about a lot in previous blogs and industry presentations—solves this problem. It’s a model where one or more parties manage data acquisition, and license access to the data to multiple clients. It allows development of a data set at scale, with both historical and near real-time data. 

It also makes this data available to projects beyond academic research, that historically have had no access at all, opening up opportunities for novel uses of the data.

MyEnergy Marketplace

Wattwatchers’ own MyEnergy Marketplace (MEM) is one such Energy Data as a Service offering. It’s a living resource for consumer energy data developed by Wattwatchers with grant-funding support from the Australian Renewable Energy Agency (ARENA). Wattwatchers now maintains the MEM as a one-stop shop for energy data. 

The MEM is a continually expanding, ethically sourced and easily shareable dataset from over 5,100 Australian sites under Wattwatchers near real-time monitoring.

We look forward to seeing many more exciting use cases for energy monitoring data like this one, that the team at UNSW have uncovered.

If you have a project that requires residential electricity data, visit the MyEnergy Marketplace website to get started.

Adriaan Stellingwerff is a Senior Software and Data Engineer at Wattwatchers.