ISA 18.2 Alarm Management in Process Plants

Alarms are a very important part of automation engineering. Alarms help in tripping a system, warning the operators, and safeguarding the system based on any critical or non-critical conditions. However, many engineers who design HMI or SCADA programs often include alarms that are not necessary. Too many alarms, incorrectly linked alarms, or poor alarm management can overwhelm the operator and affect plant performance. To address this, the International Society of Automation (ISA) developed a standard for alarm lifecycle management in process industries, called ISA-18.2. In this post, we will see the concept of ISA-18.2 alarm management in process plants.

Contents

Why is ISA 18.2 important in process industries?

As discussed earlier, we are using this standard for alarm management. First, let us understand what this concept is. Suppose you have a pressure transmitter, which has four levels of alarm – low low, low, high, and high high. You are comparing the pressure value with predefined alarm limits (e.g., Hi, HiHi, Lo, LoLo) and generating an alarm when the value breaches those limits. Now, as pressure varies, even a slight change will cause an alarm to be shown. In that case, you will have so many alarm logs that are unwanted. This brings alarm management and the ISA standard into the picture.

The aim of this standard is to reduce unwanted alarms and keep only required alarms. Alarms should be configured with time delays, deadbands, or hysteresis to prevent nuisance alarms caused by short-duration signal fluctuations. Also, it recognises that it gives priority to alarms, so that the user knows which is critical and which is not. The alarm window designed on the screen should be such that it is easily understood by the operator. And lastly, it relates to who will delete the alarm and also log the activities of the users who are operating with the alarm logs. These practices are recommended in the ISA-18.2 standard to ensure alarms are relevant, prioritized, and operator-friendly. We will discuss later in the post how it is designed.

Implementing alarm management as per the ISA 18.2 standard

ISA-18.2 defines an alarm management lifecycle, including stages such as alarm philosophy development, identification, rationalization, implementation, operation, and maintenance. which are as – development of document explaining the logic of alarms generated, categorising the alarms based on their criticality, giving correct message names to each alarms, counting the final alarm count and removing unwanted alarms, deploying the logic in PLC and configuration of the same in SCADA, and generating each alarm in runtime and logging their performance with frequency.

Now that we are clear why this standard is required, let us see some general points that comply with this standard:

Priority must be given to alarms based on their category and criticality. This alerts the operator to the required action to action.
All the alarm states – active or normal, acknowledged or unacknowledged- must be shown near the alarm triggered. Each acknowledgement must be logged in the audit, as it will show when the operator has done the action. Also, all the active time and normal time must be logged in some database for record.
Alarm counts are recommended to be shown, as this helps the operator to understand how many times this alarm is recurring.
Operators should be allowed to shelve or suppress alarms temporarily during runtime, with proper logging and access control to prevent alarm flooding and operator confusion.
Two windows are recommended to be given in graphics: alarm history and alarm summary. Alarm history shows all the history of alarms logged, whereas alarm summary shows the current active alarms. Also, an alarm banner must be shown on every screen, which helps the user to attend to the alarm immediately.
If an alarm is deemed unnecessary, it should be evaluated during alarm rationalization and removed through proper change management protocols.

Practically speaking, if you fail-safe the alarm logic in the PLC and follow the standard, then the following steps are recommended:

An alarm, on and off delay timer must be present. Alarms are triggered when the process variable crosses defined alarm limits, and hysteresis or deadbands must be used to prevent chattering near those limits.
An acknowledged alarm should remain active on screen until the condition clears. Acknowledgement should not reset an alarm if the condition still exists.

ISA 18.2 Alarm Management in Process Plants

Refer to the image above to understand the cycle of the ISA 18.2 standard more clearly. As seen and discussed before, all the stages of the alarm standard are covered and executed in a closed loop, so that the standard never fails.

In this way, we saw the concept of ISA-18.2 alarm management standard in process plants.

Read Next: