Modeling Market Incident Levels (Warranty & Over) using Weibull Distribution (Part 2) English Free

  • Category SIA Collaborative Documents
  • Edition SIA
  • Date 05/31/2022
  • Author Lavanya BONVIN & Caroline RAMUS SERMENT - Stellantis | Nicolas FORISSIER - Groupe Renault | Rémi LARONDE - Valeo
  • Language English
  • Type PDF file (1.02 Mo)
    (Downloadable immediately on receipt of online payment)
  • Number of pages 26
  • Code DC-11-01
  • Fee Free

This document has been written by a group of quality experts of  Renault Group, Stellantis and Valeo, as part of the work of the SIA Expert Community Reliability - Quality - Safety.

Modeling Market incidents is a key input

  • to master warranty costs (impact on company profit)
  • to prioritize the development and the deployment of countermeasures (impact on Customer satisfaction)

The main purpose of this document is to detail simple methodologies of some uses related to the items mentioned above.

The “part 1” document (DC-07-01) only deals with an easy way of modeling Market incidents considering

  • 1 mode / 2-parameters Weibull distribution
  • Maximum Likelihood Estimation method (MLE)

This document (part 2)

  • enlarges the scope of Market incident modeling up to 2 modes / 4 parameters
  • and also details the methodologies considering these Weibull models to build the following uses:
    • #1 - High mileage (over warranty period) reliability specifications achievement level
    • #2 - Temporal predictions of “left to fail” during warranty period

The purpose of

  • Use #1 is to assess whether or not the high mileage reliability requirements are met in the Market (considering the modeled failure distribution enabling failure proportion prediction)
  • Use #2 is to build the “monthly” volume table of the parts which can fail for a given issue (considering the modeled failure distribution, the related monthly production volume and mileage distribution)