Research on Workload-Based Prediction and Evaluation Model in Power System

摘要

In the process of power grid dispatching, inappropriate workload may reduce dispatcher's work efficiency, and even lead to accidents. The purpose of this paper is to predict the current workload level of dispatchers, evaluate the current human risk level, and design a safe and reasonable work plan. In this paper, a multi-resource occupancy model based on VACVP (Visual Auditory Cognitive Voice Psychomotor) is proposed to predict workload. Meanwhile, a comprehensive evaluation model is proposed to validate the prediction model, which mainly uses PCA method to analyze the characteristics of physiological indicators such as heart rate, voice, movement to work out the actual workload. Finally, by means of time stamp alignment method, the workload results obtained from the two models are aligned and compared. Experimental results show that the workload predicted values obtained from VACVP workload prediction model are in line with the actual workload process. Furthermore, in a certain period of time, the trend of workload forecasting value is consistent with that of actual workload value, and the average workload data error is 0.6 grade, these ensure the validity and accuracy of the workload prediction model.