Optimization of Geometric Parameters for Double-Arch Dams through Bayesian Implementation

Abstract

This paper comprises the search of the optimal geometric parameters (area and volume) for double-arch dams. The approach is structured in several consecutive stages. The process begins with a definition of the problem about Bayesian estimators, to define the dam-shape design values. After that, an iterative sequence of equations calculation was developed until reaching a solution that satisfies the a priori established constraints. A modeling of the optimized dam has been carried out to estimate static and dynamic internal stresses. Data were retrieved from inventories of existing dams, whereas to obtain the unknown data, a Gaussian distribution under hypotheses of the Bayes’ theorem has been employed. This theorem converts the a priori distribution, through unknown parameters, into the a posteriori distribution providing expected estimators. The design of the dam shape is strongly based on the experience, therefore by collecting real information about existing dams a more accurate analysis is possible.