Scheduling and Situation-Adaptive Operation for Energy Efficiency of Hot Press Forging Factory

Abstract

Hot press forging is the process of shaping heated metal into a desired configuration by applying pressure. It is a highly energy consuming process due to the need of heating the metal to a high temperature. To save the energy, we propose to optimize job dispatching policy to be used for scheduling the jobs, by searching through the policy space. In doing so, each candidate policy is to be evaluated through a simulation of applying the policy to scenarios of forging productions. For simulations fast enough to enable the search, we use predictive models for energy and time cost of each processing equipment, obtained by learning from the process data collected via IoT sensors. The dispatching policy thus obtained also enables adaptation to changing situations by being used to reschedule the jobs in a real time.