Modeling of Operator Performance for Human-in-the-loop Power Systems

Human operators interact with the power control system as "in-the-loop" control elements to ensure the system stability and safety. The role of human operators becomes more critical with the increasing usage of renewable energy resources. This research seeks to support operator training by developing a technique to quantitatively model human operators' performance in the context of a simple power dispatch task. In the designed compensatory tracking task, the operator acted on the system output error and was part of a feedback control loop. The primary metric developed to evaluate and model the operator's performance is the normalized deviation, which is defined as the difference between the individual quadratic error and the averaged performance. Twenty-three human subjects participated in the experimental study. The data collected was then used to examine the proposed modeling approach and to obtain insights into possible effects of human factors. The proposed modeling technique will enable us to evaluate and compare the operator's performance to the optimal controller designed for the same task, and to design the training program that aims to align the operator's performance closer to the optimal controller.