State Based Alarming

Operating a plant is a dynamic and constantly changing task with most alarms pertaining to the normal operating state of a piece of equipment. However, equipment often has several normal, but differing, operating states. It is a best practice that all such normal operating states should not cause alarms. Alarms should be produced only upon abnormal or unexpected events. Situations like these cause alarm floods to the panel and make it easy to miss necessary and high priority alarms and alarm system becomes a nuisance, a hindrance, or a distraction, rather than a useful tool. In order to account for these expected changing conditions, state based alarming is implemented.

State-based methodologies produce dynamic alarm configurations based upon the specific process and equipment conditions. Multiple alarm set points and priority settings are configured for appropriate alarms and enabled based on plant state. When the state changes, the system changes the alarm settings to predetermined values appropriate for the new state.

Process states can be grouped into 3 categories:

  • Planned Shutdown: when a unit, equipment or area is shut down for maintenance. Alarms should either be suppressed or limits moved to a lower value to detect when isolations have failed
  • Operating modes: batch processes, when feed conditions change, start up or with redundant equipment. These are normal conditions that are expected throughout a plant and the alarm limits should be adjusted accordingly to account for the new state
  • Consequential alarms: Such alarms can result from equipment failure that triggers a multitude of alarms down and upstream; the excess alarms should be suppressed and focus should be given to the equipment trip alarm.

When configuring the trigger conditions for each state it is important to take these factors into consideration:

  • Use 2 to 3 reliable sensors to trigger a change in state
  • Can this alarm be safely suppressed without creating a hazard in the process
  • Use sensors that do not have multiple reasons for failure

The figure below shows an example of state based alarming. When the value of the tag was less than 100, the upper and lower alarm events were suppressed (the blue dots). If the conditional alarming was not present the alarms would be triggered and appear on the panel as seen by the orange dots