Synthesis of topologically-constrained water reuse network using swarm intelligence

The sustainability of water resources is one of the major concerns of the world population. As such, industries are finding ways to minimize the water withdrawals and to reduce the water pollution through efficient use of water supplies. Process water integration has focused on the reduction of the amount of water used and water discharged. Water integration is done to determine both the minimum fresh water budget and wastewater released. The water network needed to get this minimum water usage was determined by establishing the reuse scheme, from the source streams to feed the sink streams. A procedure for designing reuse networks with topological, network complexity and stream matching constraints was developed. The procedure used particle swarm optimization (PSO) enhanced with genetic mutator. The result achieved by PSO was compared with commercial genetic algorithms (GA) package. Four main case studies with sub-cases were used to test PSO. The freshwater saving achieved by PSO and commercial GA package with the same number of function evaluations ranged from 6 - 54% and 1 - 43%, respectively. The amount of water saving varies depending on the cases conducted. PSO provided a better result than the commercial GA package. The PSO algorithm was further improved by introducing mutation to integer variables and seeding strategy