2018 年 13 巻 4 号 p. JFST0025
In this research, 2D shallow water equations are expanded by an intrusive polynomial chaos approach for efficient uncertainty quantification in tsunami inundation flows. Uncertainty propagation can be evaluated by solving the expanded 2D shallow water equations, and then probability measures such as mean, standard deviation and probability density function can be obtained for arbitrary variables. An uncertain input is given on its initial condition of water height and/or bathymetry in dam break problems as well as in Thacker’s inundation flow problem. Obtained uncertainty quantification results are compared with (exact) Monte-Carlo simulation results to validate the developed approach. Qualitative agreement can be confirmed between the Monte-Carlo simulation and the developed approach. The computational cost of the developed approach is much more inexpensive than the Monte-Carlo simulation, so that inexpensive/accurate uncertainty quantification can be realized with the developed approach. By using the obtained results, stochastic hazard map considering uncertainties can be generated which will be beneficial to minimize potential damage via inundation flows.