Atmospheric Sensing of Wildland Fire Plumes Using KHawk UAS

Accurate measurement of the 3D flow pattern within wildland fire plumes can greatly help understand the local atmosphere fire interaction and the overall fire spread process. In addition, fire plumes pose significant safety challenges to manned and unmanned aircraft during fire observation or firefighting missions. The high temperature and strong turbulence in fire plumes make it very difficult for the sensing job. In fact, very few experimental works exist for in-situ atmospheric sensing of wildland fire plumes. In recent years, small unmanned aircraft systems (UASs) start to emerge as a new platform for meteorology sensing, which were usually equipped with scientific payloads such as lidar, cameras, or flow sensors.

The objective of this paper is to introduce our results to use a KHawk UAS for in-situ atmospheric sensing of wildland fire plumes. A 5 lb KHawk UAS was sent to fly over fire plumes generated by a prescribed burn in Kansas, equipped with one RGB camera for smoke confirmation, a pitot tube for airspeed measurement, a GPS receiver for ground speed measurement, and inertial sensors for aircraft turbulence response sensing. Both the collected raw data and estimated 3D wind will be provided for the period when the UAS was flying in the fire generated plumes.

UAS Platform

KHawk Zephyr 3 UAS is a flying wing unmanned aircraft, designed and built in the Cooperative Unmanned System Laboratory (CUSL) at the University of Kansas. It is equipped with an open source PixHawk autopilot, a ublox GPS receiver, and a pitot tube. The UAS can sense the surrounding flows or wind through the dynamics pressure sensor (Pd from pitot tube) and/or the inertial measurement unit (ax/ay/az, p/q/r from vehicle response). 2D or 3D wind can be estimated using statistical estimation filters such as an extended Kalman filter (EKF).

Experiment Description & Results

A prescribed burn was held on April 15^th 2019 over a grassland at Baldwin City in Kansas, which is managed by Kansas biological survey. The land is about 400 meter long and 200 meter wide. The prescribed fire was started following the ring fire pattern. A KHawk Zephyr 3 UAS was flied over the fire line for simultaneous sensing of fire and wind. The UAS followed a racing car pattern with two straight lines and two half circles in autonomous waypoint tracking mode. During the straight leg of the flight, the KHawk UAS flied over smokes generated by the fire burning for several seconds. More than 2G vertical accelerations can be observed from 756 to 560 second in Fig. 1. And the UAS oscillate around significantly during this time period (20 degree roll and -20 degree pitch), which is shown in Fig. 1. The aerial phone taken by the KHawk UAS during the plume encounter is shown in Fig. 2 and Fig. 3. In the final version of this paper, 3D wind estimated by the onboard sensors will be provided, using an extended Kalman filter (EKF). Detailed analysis will be provided include turbulence kinetic energy and eddy dissipation rate.