Human Reliability Assessment in Control Systems

Abstract

AbstractThe article presents modeling of human reliability assessment (HRA) in control systems. The proposed approach to modeling HRA is based on an estimate of the human error probability (HEP). HEP is modeled depending on several important characteristics of a person, such as the ability to fatigue, the ability to rest, accuracy, the ability to tolerate information overload, the ability to remember information, the ability to make decisions in time-constrained conditions, and qualimetric characteristics. The study determined the required source data for HRA simulation. HRA various existing assessment methods have been studied. The relevance of the task of assessing HRA is confirmed by a large number of studies over the past 10 years. Typically, modeling is based on a Bayesian network, causal models, cognitive methods. Methods of analysis of psychological and physiological properties of a person, as well as qualimetric methods are also common. Despite the diversity of HRA models, uniform standard evaluation method does not exist. Most of the existing models are complex due to the large amount of human performance data required. On the contrary, the approach to HRA modeling proposed in this study is quite simple. The HRA mathematical model is based on the dependence of HEP on the operations performed by humans in the control system. This mathematical model for HRA assessment approved in the research and design of two large information systems. The proposed mathematical model is new. In the future we plan to increase the precision and credibility of the simulation results.KeywordsHuman reliability assessmentControl systemHuman factorReliabilityHuman error probabilitiesHuman-machine system