A hybrid event tree-genetic algorithm for improved safety and risk of independent protection layers of pressure vessels. (8th October 2010)

• Record Type: Journal Article
• Authors: Azadeh, A.; Ebrahimipour, V.; Mohammadhosseinzad, M.; Eivazy, H.
• Is Part Of: International journal of productivity and quality management. Volume 6:Number 4(2010)
• Page Start: 435
• Page End: 456
• Publication Date: 2010-10-08
• Subjects: hybrid event trees -- genetic algorithms -- safety management -- pressure vessels -- independent protection layers -- risk analysis -- chromosomes -- genes -- crossover -- fitness functions -- mean time -- failure -- MTTF -- costs -- man-hours -- repairs -- delays -- event prevention -- engineers -- managers -- productivity -- quality management; Quality control -- Management -- Periodicals; Industrial productivity -- Management -- Periodicals; 658.5005
• Languages: English
• ISSNs: 1746-6474

Abstract

This paper presents a hybrid event tree-genetic algorithm (GA) for improved safety and risk analysis of identification independent protection layers (IPLs) of the pressure vessels. GA obtains the best configuration of active IPLs used in structure of the sample vessel using chromosome gene and chromosomes which are IPLs and permutations of IPLs, respectively. Also, crossover and fitness function are the mean time to failure (MTTF) and cost of man-hours repair and severity of the delay in event prevention, respectively. By using the event tree, the cost of risk and repair is obtained. GA simulation returns improved configuration of the IPLs with respect to minimum cost obtained from fitness function. Therefore, GA identifies the best permutation of IPLs. The hybrid approach can be used to improve risk analysis and consequently can be helpful to engineers and managers to identify the best configuration.