There’s never been a better time to talk renewables, and the technology around will only speed things up manifold.
Australia is on the cusp of an energy renewables revolution, with major projects like South Australia’s Tesla battery providing a path towards a grid that is clean, efficient and inexpensive.
To make the most of this growth in renewables, energy technologies should be paired with systems of insight – including visual analytics, machine learning and data streaming – in order to reach their full potential.
According to the Clean Energy Council, renewables make up 17 per cent of energy generated in Australia. Translated, that means everything from hydro (the biggest renewable energy source) to rooftop solar, account for around 5600 megawatts of power a year.
The failure of former Prime Minister Turnbull’s proposed National Energy Guarantee cast light on both the growing market for renewables in Australia, and the ongoing barriers to increased adoption. The proposal attempted to deliver reliable power and meet Australia’s international emissions commitments.
By using analytics and machine learning, renewables companies are poised to uphold their commitments to deliver power on schedule, at the right price, and with zero emissions.
Internationally, a recent report from Citi is strongly in favour of using data, analytics and machine learning as a means of promoting the growth in renewables.
The report found that machine learning technologies have the potential to upturn the entire energy value chain, from finding energy faster and producing it more efficiently, to more customised energy at the user level.
Citi goes so far as to say that energy could eventually become uncommoditised, meaning that one day, power could be free.
But hang on, haven’t we heard that one before? Nuclear power in the 1950s was supposed to provide limitless power; power that was, in the terminology of the time, “too cheap to meter”.
Then reality set in. Nuclear power was, and remains, expensive to design and build, and the disposal of the waste products is still an intransigent problem.
Renewables, coupled with big data and advanced analytics, might change this playbook and bring the potential of cheap power to Australia sooner than we might think. Inherent to this trajectory is the use of sophisticated analytics to understand everything from the maintenance schedules of off-shore wind turbines, to predicting when the sun will shine and when the wind will blow.
Take solar energy for example. Even in sunshine-rich countries like Australia, the sun sometimes goes behind clouds, stopping the generation of energy in its tracks. Solar plants have always been equipped with sensors able to detect the weather and report on the ambient conditions. Renewable energy generation companies are starting to harvest this data and use advanced analytics to better understand and predict the weather.
In particular, with modern machine learning and artificial intelligence technology, power companies can build reliable forecasts of sunshine and energy production.
Such models enable capacity planning in alignment with forecasted demand. Coupled with solar plant production surveillance and optimisation; this enables renewable energy companies to issue power contracts guaranteeing they will produce a certain amount of power over a given time frame at a specified cost.
Similar analytics technology can be used for wind farms. Energy companies can predict when the wind will blow, turning their turbines, and can then offer assurances that they are able to deliver on the contracted amount of power.
Vestas, a global market leader in manufacturing and servicing wind turbines, has taken data analytics a step further when it comes to wind generation.
Locally, Vestas has been nominated as the preferred supplier for two new wind farm developments in Victoria, supported by the state government’s Victorian Renewable Energy Targets (VRET).
In its commitment to reducing the cost of energy for customers, Vestas uses machine learning and streaming analytics to scrutinise the massive amount of data provided by its wind turbines. The resulting insights allow Vestas to effectively manage resources across the business, from adapting to the changing needs of existing wind farms, to producing real-time risk profiles for the assessment of new opportunities.
Vestas is able to remotely monitor and control turbine operations; including temperature cycling for defrosting and even blade rotation so that the wind farm operates efficiently as a colony to efficiently capture transient wind patterns.
Analytics can also be used across the renewables sector to minimise the downtime caused by maintenance-related tasks. By remotely collecting data on the performance of a wind turbine in real-time, the facility’s owners can predict with great certainty if there’s going to be a maintenance issue.
They can then act pre-emptively to head off the mechanical problem by scheduling a replacement part at a time when the plant is not generating power, such as at night, if it’s a solar plant, or during a lull in the windy season if it’s a turbine.
This condition-based maintenance also has a net benefit to profit for the energy companies – it’s cheaper to replace a part before it fails, and have production continue unabated than it is to reactively replace a part that has failed and stops the production of energy.
Closer to our homes and workplaces, transactive energy and the commoditisation of technologies such as rooftop solar can significantly supplement, and perhaps eventually supplant, the traditional grid, and offer a massive challenge to incumbent players who are wedded to generating and distributing energy in old fashioned ways.
By using analytics and machine learning technologies, the power generation from rooftop solar, as well as domestic batteries, and the ability of electric cars to store energy when they’re not being used could lead to a grid where literally everyone is generating electricity for the community.
One can envisage how sophisticated algorithms could control the flow and supply of power generation from a myriad of sources. This could lead to a decentralised grid, where power is flowing from many directions, and sophisticated machine learning algorithms driving allocation to end users.
By building systems of insight incorporating real-time sensor data and sophisticated analytics, countries like Australia can accelerate the transformation of energy generation from fossil fuels to renewables. We might only be at 17 per cent now, but that number is climbing – and fast.