Model-based vehicular prognostics framework using Big Data architecture

Nowadays, the continuous technological advances allow designing novel Integrated Vehicle Health Management (IVHM) systems to deal with strict safety regulations in the automotive field with the aim at improving efficiency and reliability of automotive components. However, challenging issue, which arises in this domain, is handling a huge amount of data that are useful for prognostic. To this aim, in this paper we propose a cloud-based infrastructure, namely Automotive predicTOr Maintenance In Cloud (ATOMIC), for prognostic analysis that leverages Big Data technologies and mathematical models of both nominal and faulty behaviour of the automotive components to estimate on-line the End-Of-Life (EOL) and Remaining Useful Life (EUL) indicators for the automotive systems under investigation. A case study based on the Delphi DFG1596 fuel pump has been presented to evaluate the proposed prognostic method. Finally, we perform a benchmark analysis of the deployment configurations of ATOMIC architecture in terms of scalability and cost. (C) 2019 Elsevier B.V. All rights reserved.