TABLE OF CONTENTS In order to analyze and predict the performance of a system, one needs performance (i.e., failure rate) data of all the components. Where does one get this sort of information?
Hopefully, each facility has maintenance records indicating failure rates of safety related devices in the plant. The process safety management regulation (29 CFR 1910.119, Appendix C, Part 9) states that users should be keeping this sort of data. Many plants do in fact have the information, although it may not initially be collated in a usable fashion. Experienced technicians generally have a very good idea of how often they have to repair and/or replace different equipment.
Plant maintenance records are the best potential source of data. They're most representative of the application environment, maintenance practices, and vendor of concern, although obtaining statistically significant sample sizes are important.
But if one doesn't have such records, then what?
One could ask vendors, but feedback from those who have tried is not very encouraging. Note that the 1996 version of ANSI/ISA-84.01 stated the logic solver supplier was supposed to provide such data. Even if the vendor does provide data, it's important to ask how they got it. Is it based on field returns or indoor laboratory experiments? How many users of a ten-year-old piece of equipment (e.g., a transmitter or solenoid valve) actually send failed devices back to the factory?
One PLC vendor stated that their...
