Electricity demand response is considered a promising tool to balance the electricity demand and supply during peak periods. It can effectively reduce the cost of building and operating those peaking power generators that are only run a few hundred hours per year to satisfy the peak demand. The research on the electricity demand response implementation for residential and commercial building sectors has been very mature. Recently, it has also been extended to the manufacturing sector. In this paper, a simulation-based optimization method is developed to identify the optimal demand response decisions for the typical manufacturing systems with multiple machines and buffers. Different objectives, i.e. minimizing the power consumption under the constraint of system throughput, and maximize the overall earnings considering the tradeoff between power demand reduction and potential production loss, are considered. Different energy control decisions are analyzed and compared regarding the potential influence on the throughput of manufacturing system due to the different control actions adopted by throughput bottleneck machine.
Safe Reinforcement Learning-based Resilient Proactive Scheduling for a Commercial Building Considering Correlated Demand Response
