A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering, School of Engineering and Technology

Thesis (M.Eng.) - Asian Institute of Technology, 1980

Four well-known numerical methods developed for solving flow in pipe networks namely: Hardy Cross, selected loop, accelerations factor and node iteration technique are analyzed together with a numerical method proposed in the present study. No direct literature has been made to describe clearly the feasibility and applicability of each method when used in solving flow in pipe networks of different sizes and hydraulic characteristics. The present study has contributed the knowledge on the numerical characteristic of each method through the comparison of results of six different pipe networks in terms of the number of iterations and computation time elapsed. It has been discovered that for a network with pipes of uniform or nearly uniform resistance, the acceleration factor method proves to be the best. While in the network in which the pipes have highly nonuniform resistances, the selected loop method shows its superiority above the others. The combination of the selected loop method and the acceleration factor method when applied to the network with highly nonuniform resistances can further reduce the computation time and consequently number of iterations involved.