With the growing proliferation of data borne out of industrial sensors via the internet of things (IoT), and the seemingly unlimited processing power of the cloud, energy companies are seeking opportunities to leverage expanding oceans of data to improve their operations. They are increasingly investing in artificial intelligence (AI) to provide answers to many of the critical questions they face.
AI gains ground across industries
AI is already having a significant impact on many aspects of human life. Amazon’s Alexa and Google Assistant are using the power of cloud computing, big data processing, and AI to interact with millions of people – answering questions, providing directions, and controlling home appliances, lighting and air conditioning. Self-driving cars, in large part controlled by AI, have taken to the roads, and despite a few high profile mishaps, are projected to ultimately displace human drivers…for better or worse. In these consumer markets, the impacts and influences of AI continue to rapidly emerge and evolve.
In commercial and industrial markets, AI is just at the cusp of influencing the way industries operate, and how commodities are produced, traded, moved, and transformed. As the machines of commodity production and processing are increasingly connected to the cloud, the potential for AI to improve operating efficiencies of these assets has drawn attention and investment from almost all corners of the market.
AI in energy markets
Oil and gas producers are increasingly leveraging the data they generate as a by-product of their operations to better identify exploration opportunities and optimize production to ensure the greatest possible recovery of their resources. With the advent of new AI tech, including machine learning, producers are looking to AI to improve their ability to optimally locate wellsites, model well designs and reduce drilling time and costs.
ExxonMobil, for example, has partnered with MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) to create self-learning, submersible robots that act autonomously to better ensure safe and efficient offshore drilling and production operations. NVIDIA and Baker Hughes, a GE company, have partnered to use AI and GPU-accelerated computing to help the oil and gas industry leverage AI across the value chain, noting that “from seismic modeling and automated well planning to predicting machinery failure and optimizing supply chains, deep learning neural networks can unlock insights from data that were previously as hidden as the oil underground.”
The global power industries are turning to AI to help optimize an increasingly complex infrastructure of diverse generating, transmission and distribution assets. The UK’s grid operator, the National Grid, is currently integrating AI technology from Google-owned DeepMind, with the goal to improve the power network’s transmission efficiency by as much as 10%.
In the US, the Department of Energy (DOE) has made grid security and operational stability a national priority, and in doing so, has invested more than $4.5 billion via grants in a number of smart grid technologies, much of which will rely on AI. With over 15 million smart meters and millions of other field sensors, AI applications are set to improve operations and reliability by monitoring and analyzing immense data streams in real-time to better balance loads, identify faults and continuously adjust power generating and delivery assets across the country.
Demonstrating the increasing commitment to AI in the US power markets, AES Corporation announced late last year that they will begin investing in the AI technologies required to improve monitoring, maintenance and management of their various assets, including conventional and a range of renewal generation.
AI in energy trading
Energy trading is moving toward artificial intelligence to improve already existing algorithmic (algo) trading systems. Algo systems, adopted from the financial markets, are being used increasingly in real-time power markets, helping to find profit in the rapid price movements of markets that are being dominated by highly variable renewable energy sources - like those in Europe.
And now, with the use of neural networks to power AI solutions, these systems will become increasingly smart – consuming more data points and making decisions based not only on price movements, but also on a wide range of data and information, including micro weather conditions, meter level consumption data, and potentially even information derived from social media postings like twitter.
Though almost all current algo systems rely on pre-programmed algorithms (as their name implies), there clearly is an AI “arms race” shaping up as trading companies increase their investment in technologies like machine learning and predictive analytics that provide even a small advantage in what might otherwise be unprofitable trading markets.
Perhaps the best indicator of increasing interest in - and future market impacts of - AI technologies is the investments being made by “traditional” energy players in that space. Late last year, accounting firm BDO released a report on mergers and acquisitions activity involving energy companies and AI/data analytics startups. Investments by traditional energy firms had increased from $500 million in the first quarter of 2017 to $3.5 billion in the second quarter.
While similar numbers aren’t available for the last half of the year, investments in AI are clearly continuing to accelerate, both in the private and public sectors. Last month, the UK government announced they would be partnering with private companies to invest more than $1.3 billion to accelerate AI development across three industries – transportation, energy and healthcare. In the US, the Trump administration announced this month they would set-up a new task force dedicated to US artificial intelligence efforts - this after citing a 40% increase in federal government investments in research and development of AI technologies since 2015.
Given the scale of interest and investment in AI technologies, it’s a near certainty that AI will have a profound impact on energy markets. It’s only really a question how soon…and just how much influence or control over the operational energy infrastructure, and commercial markets, we are willing to cede to these new thinking machines.