| dc.description.abstract | The optimization of sensor placement in hydrogen tanks plays a vital role in ensuring efficient leak detection and risk control in hydrogen fueling stations. This thesis presents a novel methodology that aims to optimize sensor placement over hydrogen tanks of varying sizes using a genetic algorithm. The objective is to maximize the detection performance while considering factors such as tank dimensions, sensor radius, and other variables relevant to the construction of the algorithm.
The research begins by studying the physical characteristics of hydrogen, in- cluding trajectory and dispersion patterns, through the simulation of leak sce- narios using HyRAM+ software. Understanding the behavior of hydrogen dur- ing leakages is crucial for identifying optimal sensor type and positions that can effectively detect leaks and mitigate potential hazards.
To achieve optimal sensor placement, a genetic algorithm is implemented, by iteratively evaluating and evolving sensor configurations, the genetic algorithm identifies the best individual that maximizes the detection performance.
The methodology is applied to two distinct scenarios, each representing dif- ferent conditions, including pressure levels and hole sizes, which are commonly encountered in practical settings. Additionally, the methodology is extended to the Kjørbo station, which is of particular interest due to an accident that occurred in 2019 as a result of a hydrogen leakage: this specific scenario is noteworthy due to the unique characteristics of the tank sizes involved. | |
