Abstract/Summary

Given the increased international trading in ports around the world, there are significant challenges facing ports such as rising energy consumption and greenhouse emissions. The electrification of Rubber Tyred Gantry (RTG) cranes is one approach used to reduce gas emissions and fuel costs at port, but has also increased the electrical demand across the electrical distribution network. This will force port operators to reinforce the low voltage network to meet this increased demand and remain within the operating constraints. An energy storage system is one potential solution to increase the energy efficiency of the low voltage distribution networks whilst avoiding expensive reinforcement of the power system. This thesis aims to highlight and address the peak demand problem in the network of electrified RTG cranes and attempts to reduce peak demand and electricity costs by optimality controlling the energy storage system by utilising load forecasts. Since there is currently lack of understanding of the volatile demand behaviour, the research begins by investigating the unique characteristics of the electrical demand of the RTG crane. This understanding is a vital tool to develop an accurate forecast model and maximise the benefits of using an energy storage system through a control system. Several short-term load forecast models have been developed based on the ARIMAX and ANN models to predict accurate day-ahead electrical R TG crane demand. The forecast results show that the highly volatile demand behaviour creates a substantial prediction challenge compared to normal residential low voltage network demand. This thesis then presents the significance of forecasting the crane demand to improve the energy performance of an electrical distribution network with an ESS by employing several optimal controllers. The novel optimal control algorithms considered for the network of RTG cranes are split into: a Model Predictive Controller (MPC) with rolling forecast system and a Stochastic Model Predictive Controller (SMPC) based on a stochastic prediction demand model. The proposed MPC and SMPC control models are compared to an optimal controller based on a fixed load forecast profile and a common and standard set-point controller. Results show that the optimal controllers based on a load forecast have improved the storage device performance for the peak reduction and cost savings compared to the traditional control algorithm. Further improvements are then presented for the receding horizon controllers, MPC and SMPC, which better treat the volatility of the crane demand and the uncertainty in the forecasts. Furthermore, an economic analysis of the results for different ESS location scenarios is presented to assess their viability.