Everyone wants data, and generally everyone agrees high quality proven in use data is the best. Properly designed experimental testing is one means to simulate actual operation. That is what we all learned as far back as high school during our chemistry and physics labs. These tests can be set up by manufacturers or end users.
 
For end users, they need to leverage a good quality assurance plan that ensures capture of inspection, repair, and test data. A good place to start looking at how to do this is by checking into the Center for Chemical Process Safety Process Equipment Reliability Database (PERD) initiative.
 
It is a little more difficult for manufacturers. They do not operate plants that use their equipment so they have to do something that is more along the lines of simulation. At this point, it is important to note that the same equipment can be operated in several different operating modes such as, continuous, low demand protection, and cyclic operation. Just as equipment can be used in different operating modes, it can also be subjected to different ambient and process conditions. If they choose, ambient and process conditions can be simulated. When it comes to operating modes, it can be more of a challenge. The easiest test for manufacturers to simulate is cyclic service. Low demand and continuous is more difficult because their failure rates are related to time and equipment can have a useful life of many years, so achieving a statistically valid population can be quite expensive. Therefore, the test of choice is typically to do cycle testing to hurry up the process.
 
If the equipment is going to be used in high cycle service, the data from this testing is quite valuable. Unfortunately it is not appropriate for low demand or standby service as the failure mechanisms pertinent for high cycle demands is quite different than those for low demand operation. Low demand operation is typically associated with protective systems that simply stay in one position for long periods of time. For instance if, the protective equipment is tested every five years, then it could be in one position for that length of time.
 
The failure mechanisms for each of these operating modes are different, so they will have different failure rates as a function of time. Some companies have performed cyclic testing and then converted this to a failure rate for low demand service. In doing so with some self-serving assumptions that do not square with reality, they seem to have created equipment that never seems to fail and does not need to be proof-tested for the life of the facility. This could not be further from the truth. Cyclic service does not only ignore failure mechanisms like sticktion, its very operation makes that particular failure mechanism less likely as the cyclic action tends to clean the moving surfaces. Companies engaged in this practice are being academically dishonest and end users need to be cognizant of the actual quality and competence of the certifying body, as well as the equipment manufacturer’s product.