Assumed 10 Demands Per Year

During a SIF verification calculation review this week, the engineer pulled out a safety certificate with very low failure rates for a solenoid valve.  Certificate No. V139 2009 C4-4 [1] states that a solenoid valve has a dangerous failure rate of 4.57 E-10 failures per hour.  The exida safety certificate for this same solenoid valve shows a dangerous failure rate of 1.88E-7.  The ratio of these numbers is over 400X! How can this be?

The key is a statement on V139 2009 C4-4 which states “Assumed demand rate per year / 10.”  What process industry application has a demand rate of at least 10 times per year?  Every company I know of would re-design the process rather than allow that demand rate to happen.

Why is this statement there?  My best guess is that the data is based on cycle test results which assume mechanical wear-out is dominant failure mechanism.  High rates of cycling also preclude sticktion and cold-welding, two failure mechanisms included in exida analysis and very real in low demand safety applications. 

NEVER use failure data with high cycle assumptions in a low demand safety function.  The design might be 400 times unsafe. 

[1] TÜV Rheinland, V139 2009 C4-4, 2009-10-29.