GEOMETRY OPTIMIZATION OF TRIANGLE LABYRINTH SPILLWAY USING ANFIS MODELS AND GENETIC ALGORITHMS

Labyrinth spillways were considered as one of the appropriate options to pass PMF discharge where the system had difficulties. In a certain width in similar head these spillways with nonlinear crest can pass greater discharges compare to spillways with frontal crest. Since they pass large discharge under low hydraulic heads and need to smaller place in plane compare to other types of spillways, they are considered as economical structures. Therefore it is essential to apply optimum geometry with maximum passing discharge under specific hydraulic conditions with minimum construction cost. For this purpose in this research fuzzy inference system and genetic algorithms to optimize the spillway's geometry which satisfy the hydraulic conditions were used. To apply fuzzy inference system and to evaluate the coefficient based on available input - output pattern, from it ANFIS was employed. In this section in ANFIS model using experimental data input data such as angle of spillway walls along the flow (α), ratio of total head to spillway height (Ht/p), and discharge coefficient (Cd) were trained. Finally, using genetic algorithm and ANFIS model output, optimum geometry was found with defining the minimum cost function to satisfy appropriate hydraulic conditions.