Model predictive control strategy of energy-water management in urban households. (1st October 2016)

• Record Type: Journal Article
• Authors: Wanjiru, Evan M.; Zhang, Lijun; Xia, Xiaohua
• Is Part Of: Applied energy. Volume 179(2016)
• Page Start: 821
• Page End: 831
• Publication Date: 2016-10-01
• Subjects: Demand management -- Energy-water nexus -- Model predictive control -- Optimal control -- Time-of-use (TOU); Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2016.07.050
• Languages: English
• ISSNs: 0306-2619

Abstract

Highlights: Energy-water nexus optimization for urban water reliability is modeled. Closed-loop model predictive control proves to be more suitable than open loop optimal control model. Potential daily energy cost saving of 43.6% is achieved. The proposed models are suitable for all buildings and water sources. Abstract: The management of energy-water nexus in buildings is increasingly gaining attention among domestic end-users. In developing countries, potable water supply is unreliable due to increasing demand, forcing end-users to seek alternative strategies such as pumping and storage in rooftop tanks to reliably meet their water demand. However, this is at an increased cost of energy cost. In this paper, the open loop optimal control model and the closed-loop model predictive control (MPC) model, both with disturbances, are compared while minimizing the maintenance cost of the pump. The open loop optimal model is suitable in instances where only random disturbances due to measurement errors are present. However, in case the demand pattern changes for reasons such as occupancy change in the house, the closed-loop MPC model is suitable as it robustly minimizes the pumping cost while meeting the customer demand. Further, MPC proves its robustness as it is able to overcome the turnpike phenomenon. Each of these two models has their own strengths. The open loop model is cost effective and easy to implement for customers that have a steady demand pattern while the closed-loop MPC model is more robust against demand pattern changes and external disturbances. It is recommended that these two models are adopted according to the specific application.