DYNAMIC MODELING OF WATER SUPPLY NETWORK BASED ON HYBRID OPTIMIZATION ALGORITHM

A reasonably constructed dynamic model of an urban water supply network is the building block for the operation, simulation, reconstruction, expansion and planning of a water supply network. A dynamic model of the water supply network was built in Ma'anshan City, Anhui Province, China, which could offer science-based operations for attaining high efficiency, low cost water supply networks that foster both economic and social benefits. The model was designed consisting of 3 water sources, 977 pipe sections and 863 nodes. After that, a hybrid optimization algorithm combined algorithms based on a chaos optimization algorithms (COA), a genetic algorithms (GA) and a Davidon-Fletcher-Powell (DFP) variable metric algorithm was employed for the precise verification of both the nodal flow quantity and pipeline friction coefficients. The results showed: 1) that the guaranteed rate of <1.0 mH(2)O absolute error of pressure measurement was 84.62%; 2) that the guaranteed rate of <1.5 mH(2)O absolute error was 92.31%, and 3) that of <2.0 mH(2)O absolute error was 100%. For the flow quantity measurement, 1) the guaranteed rate of <5.0 L/s absolute error was 86.67%; 2) the guaranteed rate of <10.0 L/s absolute error was 93.33%, and 3) that of <20.0 L/s absolute error was 100%. Judging from the results above, the dynamic model was proved to be of sufficient precision. Then the simulation analysis was performed on the current status and the planned operation of the water supply network in Ma'anshan City, which proved the practicability of the designed model.