Vehicular pollution is one of the biggest concerns in urban areas across the globe. Several air pollution studies have been carried out based on deterministic, statistical and soft computing approaches. However, limited research has been carried out on understanding the non-linear and highly complex dispersion of vehicular pollution based on soft-computing approaches. Artificial neural networks (ANN) consisting of interconnected adaptive processing units can detect nonlinearity in incomplete or noisy datasets. This paper elaborates the methodology for developing an ANN-based vehicular pollution model. In this study, pollution model has been developed for Income Tax Office (ITO) intersection in Delhi, considering particulate matter (PM2.5), meteorological variables and traffic flow. Several combinations of models have been tried for prediction of pollutant concentration considering solely traffic variables and varying meteorological and traffic variables have been attempted. The results reveal that neural network models cannot predict satisfactorily when only traffic variables are considered but are able to do so when combination of traffic and meteorological variables are considered. The combination of temperature, wind direction, wind speed, mixing height and traffic flow gives the best model results. In addition, the results also concluded that wind speed is the most sensitive variable followed by atmospheric pressure, traffic flow and temperature in predicting PM2.5 concentration.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    An Approach for Modelling Vehicular Pollution Using Artificial Neural Networks


    Weitere Titelangaben:

    Lecture Notes in Civil Engineering


    Beteiligte:
    Dhamaniya, Ashish (Herausgeber:in) / Chand, Sai (Herausgeber:in) / Ghosh, Indrajit (Herausgeber:in) / Gupta, Naina (Autor:in) / Ram, Sewa (Autor:in)

    Kongress:

    National Conference on Recent Advances in Traffic Engineering ; 2022 ; Surat, India November 11, 2022 - November 12, 2022



    Erscheinungsdatum :

    29.10.2023


    Format / Umfang :

    15 pages





    Medientyp :

    Aufsatz/Kapitel (Buch)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch





    Modelling of Atmospheric Parameters Using Artificial Neural Networks

    Demirtas, Ozlem / Onder Efe, Mehmet | IEEE | 2019


    Diesel Engine Exhaust Emissions Modelling Using Artificial Neural Networks

    Rachid, A. / Ouladsine, M. / Ouenou Gamo, S. | SAE Technical Papers | 1999


    Vehicular pollution prediction modelling: a review of highway dispersion models

    SHARMA, N. / CHAUDHRY, K. K. / RAO, C. V. CHALAPATI | Taylor & Francis Verlag | 2004