Comparing operator reliability in analog vs. digital Human-System Interfaces: An experimental study on identification tasks
Massaiu, S. , Fernandes, A.PSAM Topical Conference 2017 Human Reliability, Quantitative Human Factors and Risk Management, Munchen, 2017-06-07--06-09
- Publ. year
- Publ. type
- This paper compares operators’ reliability on identification tasks in digital and analogue human-system interfaces (HSIs). Error rates and response times in experimental (non-nominal) conditions are compared for different basic tasks and for different HSI conditions. The tasks are identifications involved in performing operating procedures steps and other real control room tasks, such as assessing if a system is in operation or calculating the total auxiliary feed flow. Data are obtained at a plant training-simulator in which 16 operators answered 36 questions requiring identification of plant/parameters statuses on plant panels in one trial and on tablet-based digital overview-displays on a second trial. The results show that the task type is a stronger determinant of operator error rate than the HSI and that the HSI effect on the error rates was not statistically significant. The differences in error rates were significantly different across task types with observed error rates increases by a factor of 6 from the easiest to the most complex tasks. Furthermore, the relative differences were not consistent with the relative differences assumed by two known HRA taxonomies. Also, the observed error rates were much higher than expected, indicating that operator reliability on tasks taken in isolation is much higher than it is for the same tasks when performed as part of broader, meaningful activities. The study suggests re-evaluating common HRA practices such as the HEPs rankings of widely used taxonomies, the preeminent importance given to the HSI over the correct modelling of the tasks, and decomposing tasks into non-meaningful bits.