Design of water supply system from rivers using artificial intelligence to model water hammer

Abstract

Water hammer phenomena is one of the most destructive parameters for designing the water supply systems. This study aims to design sewage transfer system resistant to water hammers through the application of artificial intelligence. The intake location was selected with respect to sediment condition and permissible maximum discharge. After that the water pumping system had been designed and the water hammer phenomenon (WHP) was estimated using the Hytran model. Results showed that the combination of rough UPVC pipes and water hammering controlling facilitation can be chosen as the most proper system for designing the desired system. Also, the Hytran model results showed that the maximum pressure was occurred in 96.59 meters distance. On the other hand, using the adaptive neuro-fuzzy inference system (ANFIS) and a hybrid model, ANFIS optimized by particle swarm optimization (PSO), WHP was estimated in the pipes. The results showed that the ANFIS had better performance in estimation of WHP in the UPVC pipes while the ANFIS-PSO was found to be more suitable in metal pipes.