Blue Room - Session 5
1:00 to 2:00 p.m. Wednesday, June 14, 2017
Reliability Improvement of Coolant Joints Through Design of Experiments Using Life
While investigating fluid leaks on beaded tube joints, it became apparent that solution development would require application of Design of Experiments (DOE) to engineer a reliable and cost effective solution. Without DOE, it is often easy to overlook some important details and miss identifying significant contributing factors. Relying on intuition and engineering judgment alone can yield "feel good" solutions inadequate to address the root cause of failures. This paper explains the methodology used to develop a solution and validate the results.
When comparing current design of beaded tube joints to applicable SAE International® specifications, three factors were identified for evaluation: hose construction, bead size and clamp design. Tests were developed to focus on these factors. Two levels of hose construction (SAE J20R4 wall thickness and tolerance vs. current production), two levels of bead size (SAE recommended bead size vs. current company practice), and two levels of clamp design were included.
Due to time constraints, a fractional factorial DOE was used. The response was cycles to first failure during a thermocycling pressure test. The effectiveness of each DOE factor was evaluated using Weibull analysis. Factors that showed statistically significant improvement were evaluated based on cost of implementation versus warranty spending.
Key Words: Design of Experiments (DOE), Weibull, Reliability, Leak, B10 life
Corey Isaak, Nathan Keller and Mathew Thomas