Community energy – The data scientist’s perspective

Community energy - The data scientist's perspective

Australia’s first community mini-grid was proudly launched in Yackandandah in 2017 with the goal of becoming a totally renewable town by 2022.

Over the past two years we have observed how strong a community can be in; reducing its footprint on the environment, influencing the future of energy, creating opportunities, and tackling complex questions for a brighter future.

In February 2019, Yackandandah reached the impressive milestone of generating 1 gigawatt-hour of their own clean energy from the sun, the amount of energy it takes to run the lights at the Melbourne Cricket Ground for almost 3 years, and an emissions reduction equivalent to 240 cars taken off the road for an entire year.

Moha presenting at All Energy 2019

Last week at All Energy 2019, I had the opportunity to share thoughts and learnings from this journey with the industry through a presentation and panel discussion, part of which I would like to highlight in this article.

Yackandandah’s pioneering community energy project combines rooftop solar systems, battery storage and Mondo Ubi – the smart energy monitoring and management system which provides real-time information to all participating households on how they are generating and using energy individually and as a community. This information not only helps households track their energy usage, renewable status and progress towards the community’s 100% renewable goal, it also provides an opportunity to understand the dynamics of achieving its full potential.

Analysis of the aggregated community data confirmed the journey has not been a smooth ride. Sun exposure and energy usage is influenced by seasonal and weather factors throughout the year.

To be more precise, the community has been generating 2.96 megawatt-hours of solar energy on a daily basis in summer whereas this number is reduced to 1.1 megawatt-hours per day in winter (Figure 1). As a result, the community was more reliant on the grid to meet their energy needs during the colder, darker months, whilst in the summertime, they had an excess of ‘homegrown’ solar energy. Similarly, during spring, the solar generation by the community, on average, exceeds their energy need.

Does this mean the community was already totally renewable in spring and summertime? Well, defining “totally renewable” as the capacity to become independent from centralised grid-sourced energy, the answer is “not just yet”.

With the majority of households in the project being solar-only customers, the community exports the excess solar generation to the grid during the peak sun exposure hours but it still relies on grid energy overnight. The pattern of this interaction with grid is shown in Figure 2 below.

Having said that, the fact that Yackandandah has consistently generated more energy than it needs in summer and spring, suggests that with the right amount of storage capacity available and a load management mechanism, the energy generated by the mini-grid has the potential to get the community totally off grid-energy in those seasons.

With the plans for community-owned solar farm and additional storage capacity underway, the community is proudly on track on its totally renewable target.

 I have been privileged to be part of the team at Mondo and support Yackandandah on its rewarding journey. I look forward to continuing work with the amazing Yackandandah community for the remainder of their journey towards energy sovereignty, and in applying the learnings from this pioneering project to empower other communities in their commitment to make the planet a more sustainable, better place, for generations to come.

Mohadesh Ganji
Leading Data Scientist Mohadesh Ganji

Moha is a PhD qualified Artificial Intelligence and Data Science leader with experience in machine learning models across various industry domains such as energy, financial services, retail and telecommunications.

She ranked in the top 5 Australian analytics leaders in 2019, is an honorary fellow at The University of Melbourne and taught machine learning in graduate programs at Melbourne Business School.