ANN (Artificial Neural Network) model for daily prediction of Methane concentration

Sebastiano Trapani, Luigi Pignato, Salvatore Aronica, Vincenzo Piazza

    Research output: Contribution to conferenceOther


    According to the protocols for the reduction of the production of the gases responsible of the greenhouse effect, it makes necessary to develop systems able to forecast and to evaluate such gases. The dump of Bellolampo, situated in proximity of the urban area in Palermo, is a source of CO2 and CH4. It was officially created in the eighties and only some years later it was decided to convert it into a controlled landfill, soon becoming a catch basin of refuse for a large hinterland constituted by approximately 50 small towns. The city of Palermo has been equipped with a network of weather and air-quality stations far monitoring the main components that are responsible for atmospherlc pollution. The network consists of 10 fixed stations located both in peripheral and centrai parts of the city.The gases produced by the landfill, under particular meteorologica! conditions, reach the nearby inhabited centre, making to record many peak events, dangerous for the population health.The purpose of this research is to develop a pure predictive model, using an Artificial Neural Network (ANN), to forecast, with a lead time of a dai, the CH4 concentration in the site of Palermo and to help the administrations torationalize the preventive intervention to reduce the occurrence of this pollutant. The mode! uses an available time series which was recorded by the network of monitoring stations, on the urban area of Palermo, far the period 2003-2004. This study has some important implications for health warning systems environmental management in places with high pollution concentration
    Original languageEnglish
    Publication statusPublished - 2007

    Fingerprint Dive into the research topics of 'ANN (Artificial Neural Network) model for daily prediction of Methane concentration'. Together they form a unique fingerprint.

  • Cite this