Dynamic behavior of operating crew in complex systems

Individuals working as a team introduce team related issues such as interpersonal coordination, supervision, time, and task management which are influenced by nontechnical, cognitive and social skills of team members. In control room of complex systems and under high task load situations, teams of operators are solely responsible for the ultimate decision making and control; but incomplete information, little time, stressful conditions and lack of mutual awareness can lead to critical human errors and the aggregate impact of such errors inside communication loops can give rise to complicated failure modes. This research develops a method to explicitly model the operating crew as an interactive social unit and investigates the dynamic behavior of the team under upset situations through a simulation method. The ultimate goal is to study the effects of team factors and team dynamics on the risk of the complex system with the main focus being on team errors, associated causes and error management inside the team and the impact of such factors on team performance. Information sharing, distribution, and collection, building of shared mental models, team decision making and combined action execution by the operating crew are being systematically modeled and examined. An object based modeling methodology is applied to represent system elements and different roles and behaviors of the members of the operating team. The proposed team model is an extended version of an existing cognitive model of individual operator behavior known as IDAC (Information, Decision, and Action in Crew context) (Chang et al, 2007). Scenario generation follows typical DPRA (Dynamic Probabilistic Risk Assessment) methodologies. The method capabilities are demonstrated through building and simulating a simplified model of a steam/power generating plant. Different configurations of team characteristics and influencing factors have been simulated and the performance is compared based on observing param- ters of interest. These parameters include the quality of leadership, the quality of verbal and device-based communications from both individual and team perspectives, various team and organizational factors and the individual characteristics of each of operators. The mechanism of effect of each major category of contributing factors has been modeled and causal maps were developed to calculate the probability of failure of basic human functions dynamically and by using a number of lower level influencing factors driven from the existing literature. The effects of team factors and team dynamics on system risk and their impact on team performance have been studied through large number of simulation runs. The results are also compared with several theoretical models and empirical studies. As part of the analysis, the contributing factors have been modified over their qualitative spectrum, from one extreme level to another in order to see if the model generates intuitively anticipated results. We also developed CREWSIM, a customized library in MATLAB Simulink environment which facilitates the modeling process for similar applications of the methodology.