Optimization of gas detectors with the aid of computational fluid dynamics - Imposition of realistic restrictions

Abstract

Flammable or toxic substances are present during oil extraction, production, and refining. Therefore, control measures, protective barriers, and devices responsible for mitigating consequences are necessary. Flammable or toxic gas detectors play a fundamental role in prevention since they usually have an interface with the emergency shutdown, blowdown, ignition source control, ventilation, alarm system, and firefighting systems. However, technical standards and recommendations do not indicate the exact location or the required number of detectors. Although gas detectors are the most effective devices for gas phase releases, they detect approximately 39% of leaks. Several methodologies have been presented to optimize the position and quantity of the detectors. However, it is not guaranteed that mathematical solutions to the optimization problem can be implemented in practice. The unconstrained optimization result may indicate optimal coordinate sensors in a too high position, requiring cable support and making maintenance and calibration difficult. Also, another possibility is that the detectors are too close to the leak source. This condition could make detection difficult if the leak does not occur precisely in that direction. In order to overcome these difficulties, the present study incorporates distance restrictions from the leak point, distance restrictions from the ground and the equipment side. In other words, the solutions obtained from the optimization are intended to be feasible in the real world for the construction and assembly stage in the field. Furthermore, the present study compared the position of the detectors considering heuristic algorithm and binary integer linear programming, the position of the initial detectors with and without restrictions, spacings between the initial detectors, restrictions of proximity to the leak point, and alarm conditions by 20% and 60% of the lower flammability limit (LFL), which resulted in 60 different optimization scenarios. The main results showed that even with the restrictions imposed, it was possible to reach the minimum number of detectors compared to those without restrictions. All cases were detected when the initial matrix spacing was less than or equal to 1 m, regardless of other restrictions. For comparison, the 5 m spacing, adopted in most studies involving optimization of gas detectors, presented its best result with twice as many devices.