Fault detection and isolation are one of the most important steps in automotive diagnosis. In this work, a new OBD scheme is proposed dealing with fault detection and localization problem in diesel engine. Especially, the leak detection and characterization problem in diesel air path is studied. The proposed solution is based on the neural network trained using Levenberg-Marquardt algorithm in order to model the engine dynamics. This model is used to detect and characterize any leak occurred in intake part of the air path. The model is learned and validated using data generated by xMOD. This tool is used again for test. The effectiveness of proposed approach is illustrated in simulation when the system run on a low speed, a low load and the considered leak affecting the air path is very small.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Leaks detection and characterization in diesel air path using Levenberg-Marquardt neural networks


    Contributors:
    Benkaci, M. (author) / Hoblos, G. (author) / Ben-Cherif, K. (author)


    Publication date :

    2012-06-01


    Size :

    786520 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Leaks Detection and Characterization in Diesel Air Path Using Levenberg-Marquardt Neural Networks

    Benkaci, M. / Hoblos, G. / Bencherif, K. et al. | British Library Conference Proceedings | 2012


    Neural Network Supervised Control Based on Levenberg-Marquardt Algorithm

    Hong, Z. / Ruixiang, Z. / Tingqi, L. | British Library Online Contents | 2002



    Aerodynamic coefficients modeling using Levenberg–Marquardt algorithm and network

    Wang, Zhigang / Li, Aijun / Wang, Lihao et al. | Emerald Group Publishing | 2021


    Optimized Weight Selection in Levenberg-Marquardt Back-Propagation Using Wolf Search

    Rehman, M. Z. / Tanveer, Aisha / Nawi, Nazri Mohd et al. | British Library Conference Proceedings | 2020