In tbis study we will devise an algorithm which can effectively solve the simultaneous linear equation system for the looped channel network based on the similar procedure used for the Double Sweep Method. We also consider the calculation order of branches and the classification of channel networks as welt. | T~p chi Cc Journal of Mechanics, NCNST of Vietnam T. XIX, 1997, No 4 (53 - 63) EFFECTIVE UNSTEADYFLOW CALCULATION METHOD BY THE PREISSMANN SCHEME FOR LOOPED CHANNEL NETWORK NARITAKA KUBO Asian Institute of Technology, School of Civil Engineering, Thailand Do Huu THANH Hanoi University of o,:vil Engineering Preface Many Asian deltas located downstream of big rivers have been developed as paddy fields area, but they appear to be suffered from floods and also serious drainage problems. In these areas, tributary rivers, branch rivers and drainage channels often form networks, and flow direction in such waterways is usually affected by tide and is not consistent. In order to design flood control facilities or to improve drainage systems, such a technique is required as can exactly estimate and calculate water levels in rivers or channels. Usual runoff analysis which is carried out by water balance equation and simplified movement equation cannot deal with such complex phenomena and the unsteady flow analysis method should be utilized for such problems. There have been developed many numerical methods including the finite difference method and the finite element method for unsteady flow simulation. Although they have their own merits and demerits, the finite difference method is simple and effective, when we analyze the movement of one dimensional flow which can approximate river or channel flow. Especially, the implicit finit~ difference method, which has usually complex calculation procedure, has a high practical value, because it can take a very long time step and its computational time is much shorter than that of the explicit method. When we solve a problem by an implicit method, we must solve a. simultaneous nonlinear equation system once every time step, which means we must solve a simultaneous'linear equation system several times every time step. H we can not solve such simultaneous linear equation system effectively, the calculation time becomes enormous
