Artificial intelligence in renewables is driving autonomous micro-grids
Artificial intelligence (AI) overall and within the renewable sector has been a hot topic for some time now. It is not a new concept. In 1950, the British mathematician and computer pioneer Alan Turing declared that one day a machine would be able to duplicate human intelligence in every way. While we haven’t quite yet achieved that, we are seeing specialised AI techniques emerge in many industries.
However energy companies, including the renewable subset of the industry, are still struggling to figure out and implement AI.
As of now there are many cases of execution being conducted via ‘one-off’ solutions that solve a specific use case, such as predictive maintenance or demand forecasting.
However, since AI is at its peak ‘hype’, the realisation of its true benefit is likely well over a decade away. However, for firms that take the right approach now, implementation could be quicker.
In fact, we are already seeing examples of machine learning being used in activities such as the optimisation of the placement of arrays for efficiently obtaining wave energy or the prediction of polymer combinations to create more efficient solar panels.
It is important to understand that AI is not a technology discussion; it is a strategic business discussion and needs to focus on the industry challenges. For the renewable industry this is primarily availability and frequency. The constant challenge with renewable energy sources is the lack of abundant storage and unreliable fluctuation in weather, thus causing the need to still rely on traditional energy sources to meet consumer demand.
Perhaps the greatest potential for AI in the renewable space is in the growth of the micro-grid. Properly deployed, AI could allow a micro-grid to become autonomous and self-regulating by forecasting and balancing demand and supply within a small, controlled network. This would not only help balance the larger grid (less variability in demand) and make it more resilient but also enable optimal use of renewable energy sources thus helping to accelerate adoption.
Perhaps the greatest potential for AI in the renewable space is in the growth of the micro-grid. Properly deployed, AI could allow a micro-grid to become autonomous and self-regulating by forecasting and balancing demand and supply within a small, controlled network.
In this context, AI would allow for the optimal placement and configuration of the generating sources themselves, such as solar, wind or geothermal, by leveraging powerful algorithms to predict factors like intensity of the sun or the probability of a windy day. This can help industrial control modules make decisions about storage or adjusting solar panel angles so as to maximise or decreasepower output.
Deep learning could be used to analyse complex time series of various factors impacting demand, enabling accurate predictions of future consumption patterns. Information about probable supply and demand allows the grid to make “smart” choices, such as when to store energy and for how long, when to buy or sell from the larger grid to maximise ROI or when to turn on and off less efficient energy sources.
Imbuing a micro-grid with intelligence could lead to easier integration of alternative energy sources into the grid, but also to a greener and more sustainable world. A world powered by AI.
This column is co-authored by Morgan Eldred and Jimmy Thatcher from Digital Energy, a firm specialising in digital advisory for the energy industry.