Abstract：
As natural gas accounts for an increasing proportion of energy resources, there are more and more natural gas
supporting facilities, so the demand for emergency equipment is also increasing. When a natural gas leakage accident occurs,
the traditional detection methods not only have potential safety hazards, but the detection results are also used for post-mortem
analysis, and there is no real-time feedback of on-site results. In response to the above important problems, drones were used to
compensate for the poor safety and timeliness of traditional detection methods in this paper. Based on the Gaussian plume gas
diffusion model, the real-time leakage site can be calculated by concentration distribution, and the autonomous flight algorithm
was used to control the drone to fly along the on-site dangerous concentration surface to complete concentration surveying and
mapping or other on-site surveying and mapping tasks. The research also analyzed the influence of each influencing factor on
the gas concentration distribution in the model and compared the results. In addition, random errors were introduced to simulate
the gas diffusion distribution under real conditions. In this research, the autonomous flight algorithm of UAVs puts forward a
concept of "calculating while flying", which uses real-time calculation of the gas diffusion model to guide the flight of the UAVs.
The gas distribution will be updated and the flight route will be optimized while the UAV is still in flight. The gas diffusion
model part uses the concentration points measured in real time by drones to correct the error of initial results calculated by the
Gaussian gas diffusion model. Finally, the UAV autonomous flight control proposed in this paper was verified in an experimental
case. The UAV autonomous flight algorithm based on the gas diffusion model can realize the task of flying along the dangerous
concentration surface, and complete the specified dangerous concentration surface surveying and mapping task, which is very
significant for on-site accident guidance.
Key words：UAV, gas leakage, Gaussian model, path planning, autonomous detection
Corresponding Authors： kewin1983@126.com
Cite this article：温凯, 王伟, 谢宜峰, 韩旭. 基于气体扩散模型的天然气泄漏场景下无人机自主飞行控制算法研究. 石油科学通报, 2021, 04: 614-625 WEN Kai, WANG Wei, XIE Yifeng, HAN Xu. UAV autonomous flight control in natural gas leakage scenarios based on a gas diffusion model. Petroleum Science Bulletin, 2021, 04: 614-625. doi: 10.3969/j.issn.2096-1693.2021.04.043
