Author(s): Ernesto Arandia, Joe Naoum-Sawaya
Linked Author(s): Ernesto Arandia
Keywords: Energy management, real-time, pump scheduling, heuristics, optimal filtering
Abstract: Energy used in pumping alone can account for as much as 40% of a water utility's operational expense. Thus, an effective energy management strategy is to reduce electricity costs by optimally scheduling pump operations while using information on electricity tariffs and water network hydraulics. Real-time electricity pricing has been adopted in several countries and is increasingly becoming a policy goal as the best representation of efficient energy pricing in support of demand participation and distributed energy resources. Managing energy usage for water utilities under real-time pricing is challenging due to the volatility of energy prices. Planning the operation of pumps should rely on the most accurate forecasts of electricity tariffs using the most recent available data, which is typically updated on an hourly basis. This paper focuses first on the problem of forecasting electricity tariffs and water demands in real time using an optimal filtering technique. The paper then focuses on pump scheduling using the real-time input. The pump scheduling problem has received considerable attention and is usually formulated with the objective of minimizing energy expenditure while maintaining the system hydraulics in an acceptable range. Given the inherent nonlinearities of the system and the binary nature of the decisions, solving the problem requires computationally expensive simulations. One of the main challenges in performing pump scheduling in real time is therefore reducing the computational expense. The paper presents a methodology to combine real-time forecasting with heuristics for optimal pump scheduling. The method is applied in a case study to illustrate the challenges, potential benefits, and directions for further development in real-time energy management for water utilities
Year: 2017