trending Market Intelligence /marketintelligence/en/news-insights/trending/j3kwkbqehkh7_wptjn66ng2 content esgSubNav
In This List

From the rubble, utilities ready for the next disaster

Podcast

Next in Tech | Episode 49: Carbon reduction in cloud

Blog

Using ESG Analysis to Support a Sustainable Future

Research

US utility commissioners: Who they are and how they impact regulation

Blog

Q&A: Datacenters: Energy Hogs or Sustainability Helpers?


From the rubble, utilities ready for the next disaster

SNL Image

As of late January, nearly half-a-million people were still without power in Puerto Rico as a result of Hurricane Maria.
Source: Associated Press

As communities, businesses and local governments continue their recovery efforts from 2017's onslaught of storms, wildfires, mudslides and other calamities, electric utilities are working to improve their responses to natural disasters in the future. The efforts, which will likely reach into the billions of dollars, range from the technologically sophisticated, such as using artificial intelligence to better forecast weather-related destruction, to the mundane, like building temporary shelters where field workers can bunk down in the immediate aftermath of storms.

According to reinsurance giant AON Benfieldt, worldwide economic losses from natural disasters totaled $353 billion in 2017, 97% of which came from extreme weather events, such as the three major hurricanes that devastated parts of the U.S. and the Caribbean: Harvey, Irma and Maria.

Harvey, which made landfall in south Texas in August, caused at least $125 billion in damage, tying it with Hurricane Katrina for the costliest storm in U.S. history. Harvey produced flooding equivalent to all of the water that flows over Niagara Falls, 24 hours a day for 19 days, Julienne Sugarek, distribution services director with Houston-based utility Centerpoint Energy, said during a panel at the DistribuTECH 2018 conference in San Antonio on Jan. 24.

The storm took down 17 substations in CenterPoint's service territory and required 1.27 million service restorations. It also damaged hundreds of thousands of cars — "the largest single incident of automobile destruction in the history of mankind," said Michael Webber, a professor of mechanical engineering and deputy director of the Energy Institute at the University of Texas at Austin, during a keynote speech at DistribuTECH.

Amid the rising incidence of major weather events and other disasters, said Itron Inc. CEO Philip Mezey during his keynote at DistribuTECH, the primary challenge facing the power sector is "How do we keep the lights on?"

Predicting outages

The corollary to that question is "How do we turn the lights back on as quickly as possible after a major disaster?" The key to effective response to mega-storms like Harvey and Irma, which struck Florida two weeks after Harvey devastated Houston, is being "proactive, not reactive," said Victoria White, a staff engineer at Oncor Electric Delivery Co. LLC delivery services unit, during a DistribuTECH panel.

Better forecasts of weather patterns and the resulting damage "allow us to station personnel in locations that we think are going to get hit," White said. "It lets us be more accurate, and more confident that we can actually reduce downtime."

To increase that confidence, Oncor has developed an outage prediction model that uses historical weather and event data, along with big data technologies and artificial intelligence, to estimate damage scenarios, such as the number of events — including outages, downed poles, flooded substations and the like — that are likely to happen in a storm of a given intensity.

The model employs a variety of tools based around probability theory, including Bayesian networks, which are complex diagrams that map out cause-and-effect relationships among key variables; neural networks; and logistic regression — a statistical method for analyzing a dataset in with multiple variables that determine an outcome — using an IBM Corp. toolset called "IoT for Energy." The goal is a dashboard-based visual representation that will give district managers information on how to deploy equipment and personnel to restore service most efficiently.

The results so far have been encouraging, White said. "I can’t tell you it's 80% or 90% accurate. But it clusters around the right answer. [... Still,] we've incurred a lot of road blocks," she said. The data is messy, which is expected when trying to forecast extreme weather and its effects, and the team is still working on cleaning that data up. One lesson has been to widen the time range for outage forecasts from every hour to every four hours. "We've decreased precision to increase accuracy," White said.

SNL Image

No interpretation needed

Other utilities are in the process of developing similar programs. PG&E Corp. is deploying a near real-time dashboard for managers in order to optimize its fault location, isolation and service restoration technology. So far the company has rolled out the technology across 882 circuits, or 26% of its system, and calculates that a total of 1.1 million customers have avoided sustained outages as a result.

The goal, said Jessica Cropley, business systems analyst at BRIDGE Energy Group, which is helping PG&E implement and integrate the program, is to unify information previously spread across multiple systems and multiple databases into a unified visual platform. Previously, Cropley said, managers "were looking at all these systems screen by screen, checking circuits and devices to see what was online and what was offline. We've combined all that info into one dashboard where you have 10 to 20 screens vs. thousands of screens."

The result is called an intuitive visual representation: "The key was to make this data visual, with more color, more graphics," so that users "can see what the data is doing and what's actually happening in reality," added Cropley.

That is especially critical when a hurricane is wreaking havoc on the power grid. "In an emergency situation in a storm, the last thing you want to do is put the operator in a situation where they have to interpret the numbers," said John McDonald, head of smart grid business development for General Electric Co.'s Grid Solutions group, in an interview.

50 days out

No amount of big data analytics could have prepared a grid operator for the destructive power of Hurricane Maria, which virtually wiped out the electricity infrastructure in Puerto Rico. The recovery effort in Puerto Rico is not a matter of service restoration and repairing downed lines and damaged substations; it is rebuilding the entire grid for a territory of 3.4 million people. As of late January, more than 450,000 customers were still without power, according to the United States Army Corps of Engineers. Restoring power to Puerto Rico presents a chance to rebuild an entire grid using 21st century technology.

"The island was facing structural problems, which have been festering for years," Ignacio Alvarez, president and CEO of Puerto Rico's largest bank, Popular Inc., said on a fourth-quarter 2017 earnings call. "And we now have a unique opportunity to tackle these problems, not to go back to where we were, but to make important structural changes in areas such as energy, housing, health and education."

Driving across the island 50 days after the storm, Luis Herrera, director of business development for Mexico, Central America and the Caribbean for AES Corp., was startled by the devastation that remained. "I had to wonder why and how could we still see poles on houses, poles across roads and parking lots," Herrera said during a DistribuTECH panel. "In so many places there had been no clearing."

One reason was that all of the available resources had been concentrated on restoring transmission lines, leaving the distribution system in disarray. Another was the unprecedented magnitude of the 2017 storm season: Herrera and government officials expected some 60,000 distribution poles to be delivered and installed in the weeks after the storm. "But Maria happened after Harvey hit Texas, and after Irma hit Florida," Herrera said. "The poles needed to be imported, but they weren’t because they were being used somewhere else."

A few days after Puerto Rico Gov. Ricardo Rossello announced plans to privatize the Puerto Rico Electric Power Authority, the utility released an updated fiscal plan that calls for selling its generation assets over the next 18 months and allowing private entities to operate its transmission and distribution network. An AES spokesperson said the company would not comment on the PREPA privatization, but Herrera, in his DistribuTECH presentation, outlined AES’s vision for the future of the island’s grid.

SNL Image

A vision for the island

Through its new joint venture with Siemens AG, called Fluence, AES is delivering an energy storage system to Puerto Rico that will help stabilize frequency and control voltage on the grid — one of biggest problems the territory is facing right now, according to Herrera. AES proposes to divide the island into seven semi-autonomous "mini-grids" — scaled between microgrids and a conventional islandwide grid — linked with storm-hardened interconnections. With distributed solar plants and extensive deployments of energy storage, the plan would "bring that kind of confidence in the resiliency of the grid," Herrera said.

These mini-grids would each have their own generation capacity, about 70% of it distributed, which could be augmented by neighboring grids in the case of natural disasters. The cost of electricity on such a system, he added, could be half of the variable cost of power from the island's pre-Maria grid.

It is an innovative architecture, but given the island territory's political and fiscal troubles, there is no guarantee that it can be achieved. What is certain is that the next storm is coming, and grid operators from Puerto Rico to Puget Sound are searching for new ways to weather it.

The 2018 hurricane season in the Atlantic "is less than five months away," noted Iliana Rentz, program manager for emergency preparedness at Florida Power & Light Co., during a DistribuTECH panel. "We've got to be ready for it."