Monitoramento e modelagem de um poço de infiltração de águas pluviais em escala real e com filtro na tampa

Abstract

The infiltration well as a Best Management Practices (BMPs) stands out because it requires small areas for its implementation an it´s fully applicable to urban lots. This paper presents an experimental study of an infiltration well built in real scale at Universidade Federal de São Carlos SP with the objective of evaluation its performance as a BMP in the control of stormwater runoff and the removal of suspension materials. The soil was characterized as sand loam, and infiltration capacity between 93,98 and 19,24 mm per hour, these values were obtained thought field tests by the inverted well method and the double ring infiltrometer method respectively. The Rain Envelope Method was used for sizing, and the well was intentionally sized at odds with the standards of safety project. With the electronic equipment s and the precipitation measures were monitored 22 real rain events and simulated rain events, during 11 months of the well operation. During this period there was no record of extravasation of the well, representing an efficiency of 100% in reducing the runoff. After eight months of the well operation tests were conducted about the permeability of the geotextile, and the amount of powder material in the aggregate materials content in the well, in other words, the gravel and sand making up the filter cover, and also the qualitative parameters: apparent color and turbidity. The geotextile showed permeability reductions of 50,7%, 7,7% and 21,2% in the samples from the background, the wall sides, and the cover of the well respectively. The well sand filter cover retained 34,8% of fines, while the gravel retained 0,13% for the same period. About the well efficiency analyses in the removing the parameters apparent color and turbidity, it was showed that it wasn t characterized as a treatment unit of these parameters, this fact is due to the quality of water which represents the roof inlet structure. For modeling the Puls method, the saturated soil permeability defined as K non optimized was calculated for each monitored event (Kno). Kno was adjusted by non-linear regression (K optimized - Ko) for each monitored event, restricting the parameters based on values obtained in the field tests, thus for each monitored rain event was obtained two moldings for the constant values of K (Kno and Ko). The model representativeness was evaluated by the coefficients of determination ranging from 0,68 to 0,98 for the events modeled with values from Ko and 0,02 to 0,76 for the Kno, being the Ko most representative for the modeling of Puls.