Paper Conference

Proceedings of ASim Conference 2016: 3rd Asia conference of IBPSA-China, Japan, Korea


Model Development and Case Study of an Optimal Control Strategy of Central Ice Storage System for a District Cooling System

Sam Cox, Dongsu Kim, Heejin Cho

Abstract: Heating and cooling systems can be one of the largest contributors to peak electrical loads in a large building or multiple buildings. Ice storage is an effective mean to reduce peak energy consumption and to avoid high demand charges. A common ice storage control scheme would be to charge the storage tanks at a full capacity during cooling season and allow for the ice to take the place of cooling systems until the ice is completely consumed from the beginning of scheduled hours. Depending on weather conditions and cooling demand from a building(s), this strategy may prevent using stored energy during peak hours (i.e., the stored energy may be used up before peak hours). When an optimal control with proper load forecasting is available, the use of stored energy can be better planned to reduce peak electric demand and lower demand charge. This paper proposes a control system integrated with an artificial neural network (ANN) load prediction model to optimize the operation of central ice storage for district heating and cooling systems. A case study using data from the ice storage and central heating and cooling plant at Mississippi State University is carried out in this paper to demonstrate the feasibility of the proposed control system. 24hour-ahead central plant cooling load is forecasted using the ANN model based on forecasted weather data and past hour cooling loads. Then optimal operational schedules for the ice storage and chillers in the central plant are determined based on the predicted cooling load of the central plant. The results from this case study demonstrate that the proposed control system for ice storage systems can be effectively used to reduce peak energy consumption and to avoid high electricity rates.
Keywords: Ice storage, Peak shaving, Load prediction, Neural network, Optimization