Sterblue Blog


Automatic drone-based inspection is a game changer to prevent wildfires

Article first published on November 12st, 2018.

$292B. That was the total potential exposure of wild fires as assessed by the Insurance Information Institute in 2014 in a medium scenario.

More than 4.5 million U.S. homes were identified at high or extreme risk of wildfire, with more than 2 million in California alone. In 2017 more than 71,499 fires were triggered in the US.

Number of fires per state (source: National Interagency Fire Center)

The causes can be really diverse, ranging from lightning strikes to power lines passing by cigarettes and campfires.

Wildfires causes in California — 2015 (California Department of Forestry and Fire department — LA Time Graphics)

The most destructive fires comes from the power lines.

Most destructive causes of wildfires in California — 2015 (California Department of Forestry and Fire department (LA Time Graphics)

The causes of the November 2018 Camp Fire in Northern California are still under investigation but both human and equipment loss are already huge. Last year, among the 73% of wild fires for which the cause was determined, Cal Fire established that PG&E’s equipment was responsible for starting 16 out of the 18 fires that broke out on Oct. 8 and 9 in the Wine Country and resulted in more than 40 deaths.

How to prevent wild fires caused by power lines

Let’s go deeper into what can trigger a wildfire with power lines.

1) Large birds and raptors can be roosted on pole and towers. Their droppings can build up on insulators to the point that a flash-over between conductors and the crossarm can occur. During take-off or landing, large birds’ wings can touch two conductors simultaneously and create a short circuit

2) Small animals resting on transformers in substations or on power poles can also start fires by causing short-circuits when their bodies come into contact with both transformer bushings.

3) Many other conditions can lead to fires: vegetation in contact with the conductors, broken cross arms, damaged poles or bent brackets and braces can allow conductors to touch the ground or come into contact with each other

Cable igniting the dry vegetation

Implementation of routine line inspection procedures is necessary to mitigate the effects of faulty equipment. Utilities and fire protection agencies are both responsible for power lines inspection. Generally speaking the process is the following (it can off course evolve depending on the utility and the internal process):

  1. A helicopter flies over the grid once every one to three years with three people onboard (1 pilot, 1 photographer, 1 support). The helicopter can embed different type of sensors such as optical, infrared, lidar for different applications (equipment faults, hot spots, vegetation encroachment and digitization of the grid).
  2. As soon as the photographer sees a potential damage, a photo is taken.
  3. Once everybody is back at the office, photos are analyzed manually in order to annotate faults to generate a report in Microsoft Word/Excel or with some dedicated software integrated with utilities Geographical Information Systems (GIS).
  4. The utility company can then plan its maintenance actions depending on the emergencies.
Helicopter inspecting power lines

Some other inspection process also involves some ground based people inspecting distribution poles and rope access technician for transmission towers.

This process has the advantage of being fast (+80 miles can be done/day using helicopters) for the data capture phase but it suffers on several points:

1) Costly

Helicopter-based inspection costs more than $10,000 per day for data acquisition leading to a total cost up to +$400/mile including data processing

2) Dangerous

3) Not exhaustive

Even if lines inspectors can be exceptional at detecting faults on lines, mistake is a human feature! Some faults on the grid can be missed during the data capture phase and data analysis phase at the office when looking at the photos.

At Sterblue we truly think that we can support utility companies to prevent wildfires

Drone and artificial intelligence have reached a more mature level in the last 3 years which enables to really think out of the box.

What we believe in and deploy on field with our customers is easy to implement but changes things drastically. We offer a software suite which allows our customers to automate the whole inspection process. It all starts with an iPad app connected with off-the-shelf drones, the pilot selects the segment of power lines he wants to inspect and then the drone captures the infrastructure automatically. This piece of software is available for distribution and transmission grid.

Automatic trajectory generated around distribution grid on Sterblue web-app

Lots of pictures are taken during this process. Every single one is then uploaded to the Sterblue Cloud where it is automatically processed and analyzed thanks to dedicated artificial intelligence and technical experts. Finally a report on the status of the infrastructure is automatically generated.

Fault on concrete detected with Sterblue AI
Utilities needs to change their paradigm regarding data acquisition and stop thinking “the more image I take, the more I will have to process at the office”. Thanks to AI, processing dozen of thousands of images is no longer a burden, it only takes a few minutes and opens incredibles monitoring possibilities for now and the future.

Thanks to this new paradigm we can reach the exhaustiveness in capturing the assets and detecting faults. Such a process using automatic drone-based inspection solution also drops drastically inspection costs.

  • Utilities can prove to the insurance company that they are aware of the status of their entire assets.
  • Thanks to a standardized process to capture and process data, insightful information can emerged (when is this cable gonna break?)
  • Utilities save tons of money because even if automatic drone-based solution are less productive on a daily basis, the cost per mile drops by 50% thanks to automation brought by Sterblue software on the entire value chain.