Application of artificial neural networks in the dimensioning of retention reservoirs

Abstract

One of the essential needs for retention reservoirs is to reduce the volume of wastewater flows in sewer systems. Their main advantage is the potential to increase retention in the system, which in turn improves hydraulic safety by reducing the risk of node flooding and the emergence of the phenomenon of "urban flooding". The increasingly common use of retention reservoirs, the observed changes in the climate and the development of dedicated software tools necessitate the updating of the methods used to dimension retention reservoirs. So far, the best known procedures in this regard involve the application of analytical formulas and tools in the hydrodynamic modelling of current sewage systems. In each case the basis for the retention facility design is the evaluation of rainfall in terms of the probability of occurrence and duration that would result in a critical rainwater flow condition in the sewer system in order to define the required reservoir retention capacity. The purpose of this paper is to analyse of the feasibility of applying artificial neural networks in the preliminary estimation of the duration of critical rainfalls. Such an application of these networks is essential to the process of hydrodynamic modelling of the system and to determining the required retention capacity of the reservoir. The study used an artificial neural network model typically used as part of planning processes, as well as the Statistica software suite.