AI Trends Weekly Brief: AI in the Utility, Energy Industries Making the Power Grid the Information Grid

63

AI is projected to play a big role in all areas of the energy and utilities industries in coming years, from renewable energy sources, to management of demand to maintenance of infrastructure.

An April 2017 report from Indigo Advisory Group, which works with utilities and energy companies on strategy and innovation, outlines how AI has the potential to deliver the active management required of the power grid in the future. “Powerful intelligence will be able to balance grids, manage demand, negotiate actions, enable self-healing and facilitate a host of new products and services,” the report states.

AI will also help to analyze unstructured data, which makes up 80% of data in a typical utility organization.

Over the next decade, advances in AI, distributed ledgers via blockchain and robotics will have wide impact. For utilities, these trends combined with distributed energy resources, increases proliferation of sensors on infrastructure, behind the meter devices and demand management advances will unleash transformative use cases.

For example, devices that auto-detect demand levels on the grid and reduce power could be powered by AI, and recorded by blockchain.

In renewables management, emerging energy AI applications include wind turbine operation data, solar panel sensor data that gauges sunlight intensity. These are combined with atmospheric observations obtained by radar, satellites and ground weather stations. AI is also being applied to energy storage, such as by helping to estimate the useful lifetime of batteries.

In demand management, a series of AI platforms are in development with a focus on energy efficiency, energy performance in buildings and systems that anticipate user behavior to optimize energy consumption. Also, AI and game theory are being applied to reward/penalty mechanisms to ensure enough customers in a pool.

In infrastructure management, machine learning algorithms are being employed to analyze risks and opportunities across the infrastructure. AI methods are used to model scenarios and advise on actions and impacts. Also, AI is being deployed for operation and maintenance of generation sources such as gas turbines, to minimize emission of nitrogen oxides. Siemens for example is using a neural model that alters the distribution of fuel in a turbine’s burners to increase efficiency.

The Indigo report identified a number of startups deploying AI in energy and utilities. These include:

  • AppOrchid, deploying deep learning and natural language processing to understand grid behavior under variable wind conditions;
  • GE Energy, using AI to enhance wind turbine efficiency in Japan;
  • NexTracker, using software to increase the energy production of solar farms;
  • Smart Cloud, capturing knowledge in real time to improve management of complex situations; and
  • Verdigris, applying AI to building management.

The report captions that with AI being such a timely opportunity, “many vendors rush to be trend-compliant,” with marketing materials suggesting true AI is part of their solution. “This may not always be the case,” the report suggests, advising buyers to consider the use case activity. GE for example, is finding wind turbine power output in Japan increased by some 5% and maintenance costs being lowered by 20%.

Google, using DeepMind, was reported to have reduced its total data center power consumption by 15%, translating to hundreds of millions of dollars in savings over the next several years. Siemens, in early deployments in Asia after AI took over control of a gas turbine combustion units, nitrogen oxide levels dropped by 20%. Siemens is also looking to leverage their industrial cloud platform, MindSphere, with IBM Watson to deliver predictive and cognitive analytics.

For more information, go to Indigo Advisory Group.

 

AI, Robotics and Big Data Combine to

Address Leaking US Water Pipe Infrastructure

Water in the US is handled by some 50,000 municipal utilities responsible for managing the supply and millions of miles of piping. Some 240,000 water main breaks happen each year in the US. Utilities work hard to replace pipes that can be up to 160 years old. It’s a haphazard effort. Most of the water pipes in the US are projected to reach the end of their lifecycle by 2050. The estimated cost to replace the water pipes is $1 trillion.

Fracta, Inc. is positioning to do something about it, by combining combines machine learning software and big data to optimize water pipe replacement decisions for utilities. (Editor’s Note: The company recently changed the name of its US presence from Hi-Bot USA to Fracta, Inc.)  Takashi Kato is the co-founder, president and CEO of Fracta. Previously he was CEO of HiBot Japan. Fracta is offering a scalable solution for utilities to assess pipe replacement needs, forecast risk and control spending. The company projects it can reduce liabilities by 40% between now and 2050.

HiBot Japan has a robot heritage. Its founder Shigeo Hirose built robots assigned to help clean up the Fukushima nuclear plant after the tsunami hit in 2011. The American offshoot was spun off in San Jose i 2015, to commercialize the parent company’s technology.

Lars Stenstedt, Fracta’s co-founder and COO, told Fast Company that most communities replace their pipes based on age, at a rate of about half-a-percent a year, so that it would take 200 years to replace the entire network. In a city like San Francisco, it can cost $3 million per mile of replacement pipe.

Fracta can look at a utility’s entire collection of data, and compare existing pipe ratings to maps that include national data on soil characteristics, which is a key indicator of corrosion. Then the company adds AI into the mix to predict where pipe failures are likely to occur.

“We’ll use AI and let the data tell us what the correlations are, and what the real drivers are of leaks for a utility,” Kato stated.

Richard Dasher, an adjunct professor in electrical engineering at Stanford, who worked with Kato when he was a Stanford student, stated, “I think this could be a real game changer for municipalities across the United States. This kind of improvement in the actual data that you have is essential to developing the models for how pipes need to be replaced, preferably before they fail.”

So far the company has found that the age of pipes and their failure rate are not directly correlated. Variables include how the pipes were installed, the pipe material, traffic running overhead if the pipe is under a road, and accidents from construction equipment.

For more information, go to Fast Company or Fracta.ai.

 

Utilities Are Leveraging AI for Operations

Shawn Chandler, an IEEE Senior Member and Associate Director at Navigant Consulting, Inc., within the Energy practice, said in a recent interview with IEEE Transmitter, that the utility industry is using AI to leverage big data and draw inferences from very large data sets.

In particular, AI is being used to improve common utility operations such as:

  • reliability —self-healing grids, operations improvement and efficient use of renewable resources and energy storage
  • safety —outage prediction and outage response
  • cybersecurity of systems —threat detection and response
  • optimization —asset, maintenance, workflow and portfolio management
  • enhancements for the customer experience —faster and more intuitive interactive voice response, personalization, product and service matching

For more information, go to IEEE Transmitter.

 

BP Investing in AI to Transform Drilling Operations

British Petroleum is investing in AI’s potential, seeing its potential to transform the oil business.

Dan Walker, BP’s Group Technology leader, told OilPrice.com that BP will use AI to combine datasets (flow rates, pressures, equipment vibration) with data from the natural environment (seismic information, ocean wave height) to transform the way they run drilling operations.

BP has begun testing AI technology on a small scale with new “personality pumps” in Chicago and New York, aiming for a more interactive experience at the gas station. Customers can interact with an AI personality called “Miles,” which offers customers options such as trivia, music and creating video ecards that they can send to friends. (Hope not to be the one behind waiting for the pump to free up.)

“AI is enabling the fourth industrial revolution,” Walker stated.

For more information, go to OilPrice.com.