Track 2 Session 12

9:15 to 10:15 a.m. Friday, June 15, 2012

Automatic Identification of Parts with Degrading Reliability Performance

A critical aspect of supply chain management is managing the costs caused by parts with poor reliability performance. To do this, the reliability of thousands of parts needs to be assessed on a continuous basis. Often, part failure data is not available or not suited for analysis of all the parts or components a complex organization depends on for its operation (e.g., a manufacturing plant, fleet of vehicles, etc.). In addition, traditional reliability analysis and supporting commercial software are insufficient in analyzing the thousands of parts on a continual basis. In this presentation, we will demonstrate how logistic data (e.g., tracking part orders in an organization) can be used as a proxy for failure data and with some simple modifications to traditional reliability algorithms the thousands of parts an organization depends on can be systematically analyzed. We apply our methodology to 6,000 components on an Army weapon system and demonstrate the ability to identify specific parts that have degrading reliability.

Key Words: Screening Parts

Shawn McKay, Elvira Lorado, Amber Jaycocks and Jason Eng

The RAND Corporation

USA