How AI is changing the energy industry [5 use cases]

With fundamental shifts in the way the energy industry operates, the pressure is on to innovate and excel. Power needs to be generated faster, more efficiently and transmitted with less outages, to more people, with less environmental impact and safety incidents. The right methods must be found to manage the rising number of gigawatts needed to keep the lights on and our world running. From traditional to renewable, from old to new, from generation to transmission and distribution – we need to identify how things can simply be done better.

So what can technology do to enable this transition? Can assets be better managed, shutdowns and blackouts prevented, load forecasts and failures predicted? As the energy industry embarks on it’s journey of digital transformation, the question we’re asking is ‘what can artificial intelligence do to enable and empower the right kind of change?’. We’ve put together some of the best case studies we’ve seen on how AI is changing the game for the energy industry.

FINDING THE ROOT CAUSE OF FAILURE - IN MINUTES

Presented with a asset reliability issue pertaining to a gas compressor motor, a large multinational with an offshore platform had experienced 24 months of constant tripping. Each tripping incident incurred a 24 hours shutdown and the longest period without a tripping incident was just 14 days. Using an AI platform to process 12 months of data, it was able to successfully identify root cause of the problem in 45 minutes, identifying two main contributors to the tripping. Upon resolving the two issues, the motor ran continuously for four months before another trip incident occurred (which was due to human error, rather than machine failure). It’s estimated that millions of dollars of downtime were avoided thanks to the use of AI to enable root cause analysis and predictive analysis.

DRASTICALLY IMPROVED LOAD FORECASTING

Because of the storage limitations of electricity, for optimal efficiency it’s imperative that generation match consumption at any given time. Load forecasting is, of course, a difficult and complex task to master – but one where AI has the potential to make a significant impact. Accurate load forecasting can enable generators to optimize the mix of generating units that can serve the expected load while minimizing the production costs. A variety of external factors also impact load requirements – things like the weather, the day of the week, seasonality, or hour the of the day. Using AI and its ability to synthesize information from a variety of disparate, historical datasets to identify patterns, the ability for predictive analytics to accurately identify the next day’s load can be drastically improved.