In a modern plant, every variable that can be measured is measured. In a normal size plant we might have thousands of variables measured and communicated over the network at a relatively fast rate (sub-second to a few seconds). These  measurements are almost always stored in a database as well, albeit at a much slower rate (every few ten seconds to every few minutes). Process variable measurements that are stored in a database (or process data as it is more conveniently known), is an extensive, rich and accessible source of data.

Process data provides a comprehensive history of the plant making it a valuable resource for modeling, controller design, condition based monitoring, performance analysis and maintenance scheduling. This valuable resource, however, has not been fully utilized in the design of alarm systems. To the best our knowledge, there have been very few research studies on the use of process data for alarm design, and even fewer commercial products. In our research we investigate how process data can be used for the design of alarm limits and alarm processing techniques. These processing techniques include:

• Filters (moving average, moving variance, EWMA, CUSUM, etc.)
• Deadbands
• Delay timers (on-delay timer, off-delay timer, alarm latches)

As a result of our research, we have developed Alarm Designer, a software tool (now has been integrated in our alarm mamangement platform) that allows the user to load process data and use it to design the alarm. Alarm limits, deadbands, filters and delay-timers can be designed and simulated using the tool. It also gives a comparison between the current design and the proposed one to investigate the effect of the design. A snapshot of the tool is shown here: