Bridging the Gap: Experimental and Numerical Approaches for Gas Leakage Detection and Mitigation
Please login to view abstract download link
Real-time integration of sensor networks with mathematical models is increasingly important for developing intelligent and autonomous sensing systems [1,2]. However, accurately modeling gas propagation remains a challenging task of high interest in many technical and safety-critical applications [3–5]. In real-world applications such as large-scale infrastructures, even minor variations or unknown conditions can create gaps between simulations and actual conditions, leading to significant misinterpretations [4]. In this work, we present a modular sensor-based experimental setup and how it can help to improve the precision and robustness of numerical gas-dispersion simulations. The developed test setup consists of a flexible array of gas, temperature, humidity and pressure sensors capable of capturing information about the gas propagation in a 450-litre box. By deliberately varying key source parameters, we generate a dataset that reflects the dynamic behavior of a range of different dispersion scenarios. Our results demonstrate that the setup provides several key advantages. The simultaneous acquisition of spatial sensor gradients enables calibration of the simulation to system-specific characteristics, while comparisons across varying inlet velocities and angles allow for a quantitative assessment of the model’s robustness. Furthermore, the integration of experimental data into the calibrated simulation framework significantly reduces discrepancies between predicted and observed dispersion paths and enables probabilistic interpretation of the simulation results. Overall, this study demonstrates that coupling sensor-based experimental setups with simulations not only increases model validity but also reveals critical insights into parameter sensitivities. These insights enable targeted optimization of simulation tools for safety-engineering applications, including contaminant source detection [6] and optimized sensor placement and control strategies [7]. Finally, we outline the potential of the combined sensing and simulation technology to improve situational awareness in large-scale infrastructures such as chemical plants or large public events and to mitigate the impact of gas leakage incidents.
