I’m hoping this will become somewhat of an ongoing series of articles about lessons learned from actual public safety UAS flying with the amazing Wake Forest Fire Department.
A structure fire call came in for a home that had already been burning for some time and was fully involved when we were dispatched.
I was en route within minutes to the call in the Wake Forest Fire Department UAS Tahoe. Emergency response was nominal except for the civilian motorist who was creeping along down the shoulder and about to pass across the road I needed to turn right on. The best advice I’ve ever received about such situations is just to hold back till the errant vehicle can decide what they are doing.
As I neared the incident address and the older neighborhood, the streets began to narrow and were more crowded with parked vehicles and an influx of emergency responder apparatus.
I made a decision to not attempt to reach the incident location but set up a landing zone (LZ) on the next street over. It turned out to be a very good decision. The street I setup the LZ was not a through street, had some but less overhead obstructions, and there was little foot traffic.
Upon arrival I set out my traffic cones to block off my LZ and donned my flight suit. The flight suit is a tremendous asset. Not only does it allow me to move my phones and radios to easier to reach pockets but it identifies me to the public as an emergency responder.
I maintain the UAS in a ready state to fly. After arriving I pulled out the drawer that houses the UAS and was off the ground in a minute or less. It would have taken longer to setup the internet based UAS camera streaming for the Incident Commander to view but based on the volume of smoke I could see coming over the trees and the crackling I could hear, the fire was already substantial and time was not on my side.
Upon my initial ascent I could determine a couple of instant critical pieces of information. First, I could see the location of the incident scene. Second, I could visually tell there were no higher obstructions between me and the incident scene than I had already determined.
The onboard FLIR camera was setup with structure fire isotherms and in the DJI GO app position which meant I could only see the FLIR camera on my monitor even though both visual and FLIR video was being recorded on the SD cards onboard.
The GO app selection would be a good first choice since that would allow me to automatically orbit the UAS around the scene. That would give me additional time to then tune the FLIR camera for optimal performance.
Unfortunately once I took off and could first see over the obstructions around me it was clear the incident location was going to be challenging.
It took less than 30 seconds to get on station above the fire. You can see from the picture above how challenging the incident scene was. The structure was closely tucked into an area that was surrounded with tall trees. I was concerned about the thermal column that was being generated by the active fire. This was confirmed by looking at the initial FLIR image.
The Alpha side of the structure is on the right. Since I was unable to get a good side view of the structure I was unable to visualize data from division 1 or division 2. Most of the view of the structure was from above.
With the coverage from the tightly forested lot there was no good opportunity to utilize the orbiting strategy I had mentally planned for.
With the FLIR camera only I could see the ladder truck water stream was reaching the structure and I could see firefighters on the Bravo side deck initially in the event. It appeared as if there was active rising heat from the center of the structure but that the Bravo-Charlie corner was the hottest area of the structure. I could not determine if it was a collapsed second story or on division 2. I provided this information to Command via radio.
Here are pictures from the structure after the fire was extinguished.
Here is a straight thermal image without the isotherms from late in the event.
And with the isotherms on.
Even at the end of the event the Bravo-Charlie corner remained one of the hottest locations.
And from awhile latter and a different point of view.
1. I should have relied less on the isotherm settings and used the straight thermal image more to observe the scene initially.
2. When I was requested to search the surrounding wooded area for a potential missing resident of the home I landed to change the isotherm settings since the XT thermal camera was mounted on position 2 on the Matrice 210 and the DJI Pilot app does not allow you to change isotherms. I had to move it to position 1 and use the Go app. However, this proved to be fruitless since the surrounding area was densely foliated and the temperature of the immediate surrounding area was setting off the thermal image in the human temperature range, making the detection of a person nearly impossible. I would have been better off turning off the isotherm setting and searching the area using the straight thermal image. That should have occurred to me before I spent time landing, changing camera mount location in order to change settings with the GO app, and launching again.
In the end, the resident was located nearby and was never missing.
Things That Worked Especially Well
Having the onboard visual and FLIR camera were a tremendous benefit. With the visual camera alone it would have been very difficult and substantially riskier to attempt to visualize the scene. the FLIR camera just cuts straight through the smoke as if it is not there.
The alternate LZ was fantastic. Attempting to reach the Command vehicle would have been a waste of time and probably impossible given the street conditions close to the structure. The alternate LZ was substantially safer than almost all other options.