(317e) Optimal Design of Macroscopic Water-Energy Networks Under Uncertainty

 

Rajib Mukherjee¹, Nasreen A. Elsayed¹, Ramon Gonzalez-Bravo², Jose Maria Ponce-Ortega²,Patrick Linke³, Mahmoud El-Halwagi¹

1. Gas and Fuels Research Center, Texas A&M Engineering Experiment Station, College Station, TX 77843, USA

2. Chemical Engineering Department, Universidad Michoacana de San Nicolá s de Hidalgo, Morelia, Michoacan 58060, Mexico

3. Department of Chemical Engineering, Texas A&M University, Qatar, Education City, Doha

Abstract

The Design of water distribution network should recognize uncertainties in nodal demands, reservoir/tank levels, availability of system components and so on. The inclusion of uncertainty in the model will make the solution robust taking into consideration future projected developments. In this work, optimization of water network that involves residential use, desalination units, cooling systems and industrial processes leading to water-energy nexus, has been performed taking into consideration uncertain demands at different nodes. The model also accounts for sustainable goals through environmental objectives. The network is optimized with both centralized as well as decentralized options. Multiple water sources are also used for optimal selection including desalinated sea water. In this work, we are comparing the deterministic solution with the proposed stochastic solution where uncertain water and energy demands at different time have been used to create the probabilistic model. Excess heat from industries is used to drive water treatment technologies that include thermal membrane desalination. In solving the problem, it is assumed that in future, the demand uncertainty is known but the demand curve is unknown. The stochastic MINLP problem has been formulated as a two stage stochastic integer programming using dual decomposition with Lagrangian relaxation. The designed network is optimal over different scenarios under various conditions. The results are compared with deterministic solutions where the exact demand at different time is known a-priori. Results from the deterministic as well as a probabilistic model applied to an eco-industrial park in Qatar will be presented in this paper.