An advanced gas leakage traceability & dispersion prediction methodology using unmanned aerial vehicle

Abstract

Rapid prediction of gas leakage traceability & dispersion (GLTD) is critical for the emergency management of chemical industrial parks (CIPs). Existing GLTD prediction methods are limited by the spatial coverage and number of fixed gas sensors. It is also not possible to provide any specific information about the leak scene. Inspired by emerging drone technologies, this work proposed an advanced UAV-driven GLTD prediction methodology. The dispersion concentration field of the hazardous chemical leakage is simulated based on an improved Gaussian plume model, and then high precision and efficiency prediction is achieved by an improved sparrow search algorithm (ISSA), which can periodically update the flight strategy of UAVs. The proposed methodology is validated by a case study of a CIP in China. The results indicate that, compared with the traditional fixed sensors-driven strategy based on PSO, GA and SSA, the involvement of UAVs based on ISSA can reduce the prediction errors by 86.7%, 88.2% and 53.4%. Further analysis is conducted on the impact of algorithm parameters (UAV locations, number of UAVs, sampling frequency and step size) on search efficiency. The optimal algorithm parameters are discussed from the perspectives of cost effectiveness and computational efficiency. This work has the potentials to assist in constructing an intelligent emergency decision support system for CIPs.