StartX startup Buzz Solutions out of Stanford, California just introduced its AI solution to help utilities quickly spot powerline and grid faults so repairs can be made before wildfires start.
Their unique platform uses AI and machine vision technology to analyze millions of images of powerlines and towers from drones, helicopters, and aircraft to find dangerous faults and flaws as well as overgrown vegetation, in and around the grid infrastructure to help utilities identify problem areas and repair them before a fire starts.
This system can do the analysis at half the cost and in a fraction of the time compared to humans, hours to days not months to years.
The California Department of Forestry and Fire Protection determined that PG&E’s transmission lines were at fault for the huge Kincade Fire in Sonoma, California last year. Powerlines and equipment problems have been the cause of most recent wildfires, and the wildfire season has started up again. Spotting powerline and grid equipment flaws quickly as it gets hotter and more windy can help utilities save lives and billions of dollars.
Buzz Solutions already has pilots in the works with major utilities across the country.
Presently, power companies review the status of the power lines every year by collaborating with other organizations to capture millions of images of power lines, towers, and surrounding vegetation using drones, helicopters, and fixed-wing aircraft.
Processing these images takes six to eight months and involves linemen and engineers who manually map all that data together, looking for faults and failures and flagging them for an in-person inspection.
However, during this lengthy process, lines can easily go down, igniting wildfires and forcing shutdowns or worse.
In contrast, Buzz Solutions’ AI and machine vision technology reviews the grid inspection images captured by these varied sources which are stored in the cloud. The AI runs the images through its proprietary algorithms to detect faults on all the major components of transmission and distribution lines (see figures).
The algorithm also searches for areas where vegetation may be encroaching on the equipment and posing a fire risk. This AI analysis is done in hours or days, at half the cost of the traditional process. The utilities can then take steps to evaluate flagged images and repair or replace the equipment before they cause a fire.
“It is definitely time to move forward using AI to reduce the wildfire threat. We believe the utility industry is ready to use a far better approach to keep their equipment in good working order and keep people and property safe,” said Kaitlyn Albertoli, Co-founder and CEO of Buzz Solutions.
Buzz Solutions also provides predictive modeling and analytics from historical data, asset and fault data, and weather data to proactively determine where faults and high-risk areas will likely occur in the future.
“Our vision is to use innovative technology to safeguard our energy infrastructure and environment today and help to predict where problems will crop up in the future. This is even more important as we are seriously impacted by climate change,” added Vikhyat Chaudhry, Co-founder and CTO, COO of Buzz Solutions.
According to the Bureau of Land Management’s Western Forestry Leadership Coalition, the true cost of wildfires in the United States is a lot more complicated and involves a lot more money than we would think. There’s direct costs of fighting the fire and direct losses from the fire and smoke and water, then rehabilitation costs, indirect costs and some strange additional costs.
Wildfire costs are most easily measured when they have immediate and direct impacts. This category includes federal, state, and local fire suppression costs. These costs, in turn, can be broken down into expenditures on aviation, engines, firefighting crews, and agency personnel.
In addition to suppression costs, other direct costs include private property losses (insured and uninsured), damage to utility lines, damage to recreation facilities, loss of timber resources, and aid to evacuated residents. Most of these costs are incurred during or immediately following the fire.
Immediate emergency rehabilitation costs are incurred in the days, weeks, and months following the fire and are clearly attributable to the wildfire itself.
Longer-term rehabilitation costs are harder to measure. Damaged watersheds can take many years to recover and require significant restoration activities. Post-fire flooding events can cause additional damage to the already scarred landscape, and subsequent impacts may include an increase in invasive species and heavy soil erosion.
Indirect wildfire costs include lost tax revenues such as sales and county taxes, as well as business revenue and property losses that accumulate over time. For example, properties that escape damage in the fire may still experience dramatic drops in value as the area recovers. These indirect costs are sometimes labeled as impact costs.
Additional costs, sometimes called special costs, include things such as the value of a human life. While the EPA puts the value of a human life at $7 million, the health care industry pays out an average of $316,000 over an average life which is what they consider the average value of a human life in America.
Loss of civilian life, ongoing health problems for the young and old, those with weak respiratory or immune systems, and mental health needs all fall into this category but are rarely quantified. The extensive loss of aesthetic and scenic beauty, wildlife existence, and others are also hard to quantify.
Synthesis of case studies in the BLM report reveals a range of total wildfire costs anywhere from 2 to 30 times greater than the reported suppression costs.
So the massive Kincade Fire that burned 78,000 acres and caused 200,000 people to evacuate and reportedly cost $725 million, more likely cost many billions of dollars when all is said and done.
This is an area where technology really does make a difference, and this new AI solution is an important part of that.