How drones could prevent utility equipment from starting forest fires
The 2018 Camp Fire forest fires resulted in over $16 billion in damage, claimed 85 lives, and was recorded as the 13th deadliest wildfire in California’s state history.
According to state investigators, the fire started when a hook on a Pacific Gas & Electric (PG&E) electrical transmission tower (Tower 27/222) broke during heavy winds, dropping a wire which threw sparks into the dry brush below.
PG&E has received a lot of flak for failing to take basic precautions which would have likely prevented this catastrophe.
What do you mean?
For starters, the Caribou-Palermo transmission line – of which Tower 27/222 was a part – was originally built in 1921, meaning it was around 97 years old in 2018.
Despite owning the line since 1930, PG&E had not replaced much of the original hardware. State regulators claim that PG&E’s inspection and maintenance plan for the transmission line prior to the fire was “ineffective,” and a report by the Butte County District Attorney called PG&E’s reliance on outdated and under-inspected equipment “negligent and reckless.”
An obvious way for PG&E to protect its credibility, moving forward, would be to make a notable improvement in its inspection protocol.
I’m guessing this snapshot is going to have something to do with drone inspections.
That’s right.
But, let’s do an overview of what’s being inspected first.
PG&E operates over 5,000 miles of high-voltage wires in Califirnia’s drought-prone forests. To make matters worse, the Wall Street Journal analyzed PG&E’s 20 “worst performing lines,” and found that 16 of them are in high-risk fire areas.
The combination of failure-prone transmission lines in fire-prone areas is thought to be a significant and unacceptable hazard, and a recipe for a repeat of 2018.
Part of the solution is more frequent inspections, but many lines and towers are hard to reach, making those inspections and resulting repairs difficult. That’s where drones come in.
PG&E is upgrading its system inspections program by using drones, computer vision, and machine learning, to better detect problems before another fire is started.
How do drone inspection solutions work?
In a nutshell, drones are deployed to gather data such as thermal imaging, LiDAR, and sometimes multispectral imaging, for critical infrastructure, such as distribution poles and transmission towers.
This generates terabytes and terabytes of data.
Computer vision and machine learning are then trained to classify this data and identify subtle deviations from the norm, such as the beginnings of corrosion or other damage to components.
This helps utility companies prioritize which components to replace or keep a close eye on before things get out of hand, as they did in the case of the Camp Fire.
How far along is PG&E’s program?
PG&E’s program — which has been in development since at least 2016 — is currently being used to predict how transmission equipment will handle high-wind events, to help operations staff prioritize maintenance work, and to help PG&E leadership decide whether to shut off power to a high-risk area during severe weather conditions.
Although high fire-risk areas are a top priority, the plan is to expand to lower risk areas and inspect over 15,000 miles of electrical lines in 2020.
What are the main benefits of a drone inspection program like this?
There are many.
First, these types of programs can relieve inspectors and electrical workers of routine tasks. For instance, during the inspection process, a program like the one described can alleviate the need to scan hundreds of images of each structure in a high-fire-risk area to find a right-of-way or access path for maintenance and repair workers.
This allows for more focus to be put on identifying and mitigating fire risks.
Second, once the computer vision and machine learning are adequately trained, these programs can reduce human error and speed up response time when issues are found.
Finally, by reducing the need for manned inspections, drone inspection programs can reduce safety hazards for the inspection crews of utilities companies like PG&E. (For more information on how drones improve infrastructure inspections, see our article, “Three Reasons Drones Improve Infrastructure Inspections.”)
Will this program prevent future forest fires?
The California Public Utilities Commission (CPUC) estimates that about 10 percent of the state’s wildfires are triggered by power lines.
A 10% reduction in wildfires would be a notable accomplishment for California, especially given the large scale of recent incidents, such as the Camp Fire.
(It’s also worth noting that federal investigators are looking into whether the more recent Bobcat Fire could have been caused by another utility company’s faulty equipment.)
The good news is that PG&E’s case seems to be inspiring other utility companies to develop similar drone programs. At the end of 2019, San Diego Gas & Electric started using drones and computer vision to inspect its distribution equipment in high-risk areas.
And similarly, in March of 2020, Southern California Edison announced that it was piloting a program that uses drones to inspect distribution and transmission lines in high-risk fire areas.
But unfortunately, better inspections by themselves will not eliminate the risk of utility-driven forest fires.
California has 25,526 miles of higher voltage transmission lines, and 239,557 miles of distribution lines, two-thirds of which are overhead, according to the CPUC.
Since it’s so hard to maintain and inspect overhead lines – especially old ones – many have suggested that the only truly effective way to prevent future forest fires is to move the lines underground, where harsh weather conditions are less likely to cause sparks.
But, according to PG&E’s website — Facts about Undergrounding Power Lines — it costs about $3 million per mile to convert underground electric distribution lines from overhead lines, and it costs $800,000 to build a mile of new overhead line.
So, until residents of California are willing to experience a massive hike in electricity costs, it’s likely that drones will be critical to minimizing the risks brought about by California’s aging electrical infrastructure.
Miriam Hinthorn - Contributing Author
Miriam Hinthorn is an experienced management professional who is currently pursuing her master’s in Data, Economics, and Development Policy at MIT while serving as principal consultant at Consult92.
Miriam developed a love for UAS technology when she served as operations manager at Consortiq. Today, having completed over 30 successful projects in 10 countries, she loves solving a wide variety of logistical, technical, and cultural challenges for her clients so that they can focus on what care about most.
