Human reliability analysis technique selection for life support systems maintenance of orbital space stations using Fuzzy AHP and ANN

Abstract

A methodology for assessment of a suitable Human Reliability Analysis (HRA) technique for Orbital Space Station (OSS) Environment Control and Life Support System (ECLSS) maintenance considering the ones proposed from other industries is presented. This paper presents a hybrid model using Fuzzy analytic hierarchy process (Fuzzy AHP) and artificial neural networks (ANNs) theory to develop a suitable HRA methodology for OSS ECLSS maintenance. The model consists of two modules: Module 1 applies Fuzzy AHP using pair wise comparison of criteria for existing methods as proposed from other industrial domains. Four criteria that are decided are adequacy, costs, effectiveness and efficacy. The four HRA methodologies that are considered to be evaluated are THERP, CREAM, NARA and SPAR-H. Module 2 utilizes the results of Fuzzy AHP decision matrix into Artificial Neural Network (ANN) model for ranking of all selected HRA methodologies. The results yield the best HRA model for OSS ECLSS maintenance with appropriate scores to compare the performance of each HRA model.