Artificial intelligence may have exploded in the public consciousness thanks to a new generation of attention-grabbing chatbots, but AI and machine learning have been with us for several years \u2013 and these more established models are already delivering significant returns.<\/p>\n
Electricity grids and AI, for instance, are a match made in heaven. Grids are complicated, have vast numbers of inputs and outputs, require constant oversight and create huge quantities of data. Dealing with enormous numbers of variables and complex systems is one of the things AI does best \u2013 better, arguably, than pretending to be human.<\/p>\n
Iberdrola, Europe\u2019s largest utility company, which owns ScottishPower in the UK, has been using AI for more than 10 years<\/a> to make predictions, optimise processes and detect patterns, which can be used to improve its operations. One example of this is providing customers with accurate estimates of outage times after incidents, via an app. Another is allowing engineers to plan ahead by telling them which grids or transformation centres will need partial replacements the following year, using myriad variables that predict possible issues based on historical data.<\/p>\n In a slightly more sci-fi vein, Avangrid, Iberdrola\u2019s US subsidiary, announced last year that it had partnered with Levatas and Boston Dynamics to deploy a robot \u201cdog\u201d<\/a> to improve substation inspections. The dog, nicknamed Spot, has sophisticated imaging and thermal technology and uses AI to detect damaged equipment. Automated inspections will take place more frequently and the result will be greater reliability for customers.<\/p>\n Maximising renewables output<\/p>\n But AI isn\u2019t just about making grids more operationally efficient. It is also helping to make energy more clean. Renewable sources of energy such as wind and solar are by their nature variable, but AI can help operators maximise the electricity they produce, reduce wasted energy and limit the need for fossil fuels as back-up sources. \u201cBy predicting when and where energy will be needed, AI can help avoid outages and cut down on wasted power,\u201d says Justin Bates, co-founder of ESGmark, an organisation that promotes and recognises environmental, social and governance standards in business.<\/p>\n For companies such as Iberdrola, this begins in the design phase. It might start with AI-driven models finding the optimum location of a wind turbine, based on years of meteorological data. It could then go into operations and maintenance. This may include using algorithms to predict maintenance issues before they occur, or anticipating wind or solar production for each hour of the day. Knowing what is being produced and where can help balance the grid, either by shifting energy around, by releasing stored energy (such as pumped hydro) or switching on back-up energy (which may come from fuels such as gas) where needed as a last resort.<\/p>\n Improved energy storage solutions<\/p>\n Back-up fossil fuel capacity should be needed less in the future thanks to more innovative and dependable renewable-energy storage solutions \u2013 and, again, AI is playing a crucial role in these. Last year, Iberdrola partnered with Multiverse Computing, a global leader in quantum computing solutions, to deliver a pilot project<\/a> in northern Spain to optimise the installation of grid-scale batteries. Multiverse\u2019s solution uses quantum and quantum-inspired algorithms to select the optimal number, type and locations of batteries on the network \u2013 thereby reducing the costs of adding batteries to the grid and improving network performance.<\/p>\n