Analyzing Repaired Spares
Abstract
Chris and Fred discuss how to analyze reliability data for repairable spares—and why the “as good as new” assumption can be dangerously misleading. They explore how maintenance itself can introduce new failures and distort your analysis.
Key Points
In this episode, Chris and Fred respond to a listener question about how to analyze failure data for repairable units being cycled through a system. The goal is to predict reliability and enable proactive replacement during planned downtime. However, the challenge quickly becomes more complex when repaired units are reintroduced into service. With limited data and unclear failure mechanisms, the discussion shifts from statistical methods to a more practical question: are you even analyzing the right thing? This episode helps listeners rethink how they approach reliability data when maintenance actions are part of the system.
Topics include:
- Maintenance can increase failure rates—at least temporarily. Every time you repair or service a system, you introduce new failure opportunities. This “Waddington Effect” means repaired units are often worse than new—at least initially.
- “As good as new” vs “as bad as old” is the wrong starting point. Real systems don’t behave cleanly. Repaired units may sit somewhere in between—and often behave unpredictably due to maintenance-induced variability.
- Small datasets don’t justify complex statistical models. With only a handful of failures, methods like Weibull analysis or bootstrapping can give misleading confidence. You can generate answers—but not necessarily insight.
- Failure mechanisms matter more than distributions. If failures are driven by different root causes, fitting a single statistical model won’t help you make better decisions. Understanding the physics of failure is more valuable.
- Sometimes the best move is to stop analyzing and start investigating. When data is limited, root cause analysis and direct observation often outperform statistical modeling in improving reliability outcomes.
Enjoy an episode of Speaking of Reliability. Where you can join friends as they discuss reliability topics. Join us as we discuss topics ranging from design for reliability techniques to field data analysis approaches.
Leave a Reply