Prediction of Failure Candidates of Technical Products in the Field Based on a Multivariate Usage Profile Using Machine Learning Algorithms Regarding Operating Data

  • Sebastian SochackiEmail author
  • Fabian Reinecke
  • Stefan Bracke
Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 1026)


In this paper an approach is presented, which gives the possibility to predict the probability for each specific product in a fleet to be affected by one or more failure modes over their lifetime. The approach considers the specific usage by the end consumer, which can be described by several usage variables. This gives the possibility to determine multivariate field load profiles within the product fleet. In addition, products, which are located in a usage area in the field, where several types of failure are likely to occur are assessed by corresponding probabilities of occurrence. Furthermore, it is possible to estimate the future risk probability for each product at the current field time as well as at upcoming field times. The approach is based on machine learning methods and is applied within a case study, which refers to a generated synthetic data set regarding the automotive industry.


Product fleet Failure candidates Machine learning methods Pattern recognition Failure recognition Field action strategy 


  1. 1.
    Sochacki, S., Reinecke, F., Bracke, S.: Ansatz zur Anpassung von Wartungs- und Instandhaltungspaketen auf Basis maschineller Lernalgorithmen im Hinblick auf den zuverlässigen Betrieb technisch komplexer Produkte. In: Gesellschaft für Qualitätswissenschaft e.V. (2018, in press)Google Scholar
  2. 2.
    Friedman, J.H.: Stochastic gradient boosting. Comput. Stat. Data Anal. 38(4), 367–378 (2002)MathSciNetCrossRefGoogle Scholar
  3. 3.
    Breiman, L.: Classification and Regression Trees. Chapman & Hall, Boca Raton (1998)Google Scholar
  4. 4.
    Hastie, T., Tibshirani, R., Friedman, J.H.: The elements of statistical learning: data mining, inference, and prediction, 12th edn. Springer, New York (2017)zbMATHGoogle Scholar
  5. 5.
    Awad, M., Khanna, R.: Efficient Learning Machines. Apress, Berkeley (2015)CrossRefGoogle Scholar
  6. 6.
    Lin, L., Yue, W., Mao, Y.: Multiclass image classification based on fast stochastic gradient boosting. Informatica 3 (2014)Google Scholar
  7. 7.
    Bishop, C.M.: Pattern Recognition and Machine Learning, 8th edn. Springer, New York (2009)zbMATHGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  • Sebastian Sochacki
    • 1
    Email author
  • Fabian Reinecke
    • 1
  • Stefan Bracke
    • 1
  1. 1.Chair of Reliability Engineering and Risk AnalyticsUniversity of WuppertalWuppertalGermany

Personalised recommendations