There’s a dire need to speed the planet’s shift to clean energy – and the power of Artificial Intelligence can help.
The world has gone through a number of energy transformations – from wood to coal, then to oil, gas and (partly) nuclear. These shifts were gradual and contingent on economic conditions.
Now major efforts are under way to reform the global energy sector to make it low-carbon, non-nuclear and climate-compatible. But, unlike the previous transformations, the ongoing restructuring process is driven by elevated awareness of the disastrous consequences of climate change.
The need to hasten the pace of transformation
Notwithstanding the global efforts made to revolutionise the energy business (to make it capable of coping with the variability inherent in most renewable energy generation technologies), there is still a dire need to speed up the shift to clean energy solutions. This is because, although the costs associated with green energy technologies – particularly solar, wind and battery energy storage systems – have been tumbling in recent years, the current penetration level of renewable energy systems (RESs) is on a global scale well below what is required to curb the global temperature rise to 1.5 °C, as recently reported by the International Energy Agency (IEA).
Challenges associated with modelling of RESs
One of the central reasons RESs are still considered cost-prohibitive options is due to the inefficiencies in the conventional models commonly used to analyse them. In general terms, such inefficiencies could stem from two sources: 1) improper calculation of the capacity of the renewable energy equipment that needs to be installed to supply a certain amount of electrical power to the target customers and 2) inaccurate forecasting of either the output power from renewables or the load demand for electricity.
In contrast to traditional dispatchable (controllable) power plants that generate electricity on demand, RESs must be equipped with high-priced energy storage technologies – with associated costs typically making up 40 to 60 percent of a RES project’s costs – to compensate for the fluctuations of weather-driven renewable energy sources. Owing to the fact RESs are made up of multiple components (such as photovoltaic panels, wind turbines, micro-hydro plants, batteries, power converters and so forth), the search for the optimum sizes of different components (which are highly interdependent) has to be conducted in a multi-dimensional design space. This is exceedingly complex and beyond the ability of business-as-usual kinds of optimisation methods to solve.
At the same time, preliminary design and feasibility studies of RESs use forecasted data on long-term weather patterns and energy consumption behaviours. Owing to the complicated and convoluted nature of making long-term meteorological and load forecasts, the actual investments required might be drastically different from the techno-economic assessments carried out in the system-architecting phase. Another issue is the lack of consistent historical demand and/or climatological data in remote sites where electrification has to be started from scratch. As one would expect, these weaknesses undermine the validity of assessments by system designers and consequently impair the reliability of such systems. This in turn would discourage utilities from reshaping their energy infrastructure and could demotivate consumers who aspire to become producers of energy (or prosumers).
How can artificial intelligence help?
Artificial intelligence (AI) focuses on building intelligent machines/systems that match or surpass the human brain’s abilities. In fact, AI is a broad term encompassing a wide variety of techniques, methods and approaches that help achieve this goal. Evolutionary computation (EC) and machine learning (ML) are two core sub-disciplines of AI that offer promising prospects to realise a rapid shift to a low-carbon energy regime.
EC techniques are inspired by, and mimic, the swarming behaviour of creatures and biological evolution-related mechanisms in nature (such as bird flocking, fish schooling, natural selection and so forth) and are aimed at finding the globally optimum solutions to complex optimisation problems. Equally important, ML techniques provide a platform to develop computer tools able to undertake specific tasks based on insights gained from historical data – such as weather-pattern and demand-behaviour recognition in the context of the economic planning of RESs – without being programmed explicitly.
EC techniques could be deployed to effectively solve intricate RES financial planning problems that are subject to extremely large design spaces and previously believed to be impossible to solve with high accuracy. ML methods, on the other hand, enable us to reap the full capacity of renewable energy generation equipment (through minimising dumped renewable energy) by precisely predicting the long-term profiles for climatic and energy demand data that will help match supply and demand in a cost-optimal way. Such long-run predictability offered by ML approaches can make a substantial contribution to improving the robustness, resilience, reliability and affordability of RESs.
In these ways, techniques taken from the steadily evolving world of AI (termed the “AI spring”) – as highly effective, complementary tools for the ever-falling costs of renewable energy generation technologies – can pave the way towards a more sustainable future.