Multi-Scenario Regression Prediction on Temperature of Molten Iron in Transportation Based on Cyber-Physical Energy Systems

Abstract

Although the CPS has been used in some actual industrial enterprises for some time, the prediction-based energy scheduling is a challenging and unsolved problem due to the specific restrictions of different operating procedures including imperfect measurement hardware, the complex chemical and physical reactions, manual operation and weather effects. In this paper, aiming at the optimization of the steel enterprises' energy scheduling system and fulfilling the online prediction of temperature in hot metal transportation based on CPS, we first clean and classify the data in combination with the process characteristics. Then based on the data features, we propose a multi-scenario multivariate linear regression prediction method based on prediction error. The classification method was LS-SVM method with RBF kernel. Finally, we conducted offline experiments based on the actual field data to verify the effectiveness of our method according to the hit rate within 10 °C.