A performance indicator of the effectiveness of human-machine interfaces for nuclear power plants

Effective interfaces must call up operators' deep understanding of plant operation if operators are to deal effectively with normal operation and diagnosis of transients. The present research examines the ability of a memory recall task to indicate the ability of an interface to couple plant state to operator knowledge. Novices, people with intermediate experience, and experienced nuclear power plant operators viewed three kinds of displays. They watched nine simulated transients and tried to recall the values of variables, or the states through which the plant passed, and to detect and diagnose the nature of the transients. The displays were simulated analog instruments, simulated analog with pressure-temperature graphics, and an animated representation of the Rankine cycle. The recall tasks did not show promise as indirect performance indicators of the quality of the interfaces, but the diagnosis test detected differences in the quality of the displays and the levels of expertise.