A fuzzy multi-objective optimisation model of risk-based gas detector placement methodology for explosion protection in oil and gas facilities
A flammable gas detection system is one of the critical control strategies of catastrophic events such as fire and explosion. While gas detector technology has improved significantly, adopting a methodology for optimal placement of gas detectors is still an issue, especially when integrated with a r...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Published: |
Institution of Chemical Engineers
2022
|
| Online Access: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85127211491&doi=10.1016%2fj.psep.2022.03.001&partnerID=40&md5=44fde599568ca592504ac7707015e18b http://eprints.utp.edu.my/33120/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | A flammable gas detection system is one of the critical control strategies of catastrophic events such as fire and explosion. While gas detector technology has improved significantly, adopting a methodology for optimal placement of gas detectors is still an issue, especially when integrated with a risk-based approach. An enhancement of a risk-based approach is proposed to optimise the placement of flammable gas detectors by integrating a formulation of fuzzy multi-objective mixed-integer linear programming with the goal of minimising the residual risk and total number of detectors for effective explosion protection. The proposed methodology primarily begins with the identification of critical leak scenarios that require detection followed by the prediction of a targeted gas cloud and dispersion analysis using a computational fluid dynamic model. Risk analysis is conducted to identify high risk areas that need flammable gas detectors protection, which is the input for the mathematical model. The proposed risk-based model was tested using a case study involving a natural gas liquids (NGL) recovery unit, and the results were compared to a published greedy algorithm (GA) formulation. By using mixed-integer linear programming (MILP) formulation, the number of detectors needed are lower with higher risk reductions compared to the GA formulation. Additionally, a sensitivity analysis was performed to determine the proposed model's response to parameter variations. © 2022 The Institution of Chemical Engineers |
|---|