Optimal Load Management in a Shipyard Drydock

Abstract

With the proliferation of enabling smart grid technologies, industries are increasingly looking to reduce their electricity costs through the adoption of renewable energy sources and efficient load management strategies. In this context, the realization of lower electricity costs requires the design of efficient energy management systems (EMSs). An EMS needs to factor in the unique operational requirements of the industry for which it is designed. Consequently, there has been a lot of research interest in designing EMSs for various industrial applications. However, the existing EMS formulations and models are not suitable for a shipyard drydock. Shipyard drydocks may be treated as grid-connected microgrids containing pump loads, interruptible loads, critical loads and heterogeneous generation sources. This paper proposes three modules, which can constitute a shipyard drydock energy management system (SEMS). A load forecasting module generates short term and medium term load forecasts using historical load demand data and ship arrival schedules as inputs. A multilayer feed-forward neural network with backpropagation is used to perform the load forecasting. A contracted capacity optimization module uses the medium term load forecast to find the optimal contracted capacity for the drydock. The short term load forecast and the optimal contracted capacity are used by an optimal scheduling module to reduce the total electricity cost incurred by the shipyard drydock. Case studies performed using data from a local shipyard demonstrate the efficacies of the proposed load forecasting and optimal scheduling modules in improving the accuracies of the load forecasts and reducing the overall electricity cost, respectively.