I have referenced these writers before when discussing process safety and once again they communicated the message better than I could regarding how we should analyze “process alarms” when we claim them to be a layer of protection. We should only accept alarms that have been established using a RAGAGEP in their design (audible, visual, human machine interface (HMI), and response times). Here is their more in-depth analysis…