Case Study

Alarm Management Solution

exida worked with Ashland to Implement an Alarm Management Program that Makes the Alarm System More Useful during an Upset and Addresses the Loss of Experienced Operators.


Customer:

Ashland BDO Manufacturing

Challenge:

Improve Alarm system performance to better support operations

Solution:

Develop an alarm philosophy and perform alarm rationalization

Results:

Unnecessary alarms were eliminated and alarms were prioritized to re ect urgency of response and severity of the consequences. The cause, consequence, corrective action, and time to respond were documented for each alarm and displayed via DeltaV Alarm Help.


Project Summary

Ashland ISP Lima, located next to a large Husky Refinery in Lima, OH, was experiencing alarm management issues such as alarm f oods and nuisance alarms. The BDO plant, which had been in operation since 2000, utilizes a DeltaV distributed control system. Rather than following a “DIY” approach, Ashland brought in exida to help them design and implement an effective alarm management program which included the following: 

SOFTWARE TOOLS SERVICES
  • DeltaV Analyze
  • DeltaV Alarm Help
  • PPCL CVE - Operating envelope and recommended alarm limits
  • SILAlarm - Alarm Rationalization and Master Alarm Database
  • Training on Alarm Management Practices & Principles
  • Alarm Philosophy Development
  • SILAlarm Product Training
  • Alarm Rationalization Kickoff and Facilitation training

Challenges

Alarm system performance was negatively impacting operations. After a plant upset, operators were often flooded with alarms (2000+ in 60 mins), such that the DCS (alarm system) did little to help guide the operator's response to the situation. In fact, some of the older operators wanted their PANALARM® (annunciator panel) back.  Alarm priority was not meaningful, making it hard for operators to know which alarm was most important to respond to first.

The operations team consisted of senior operators that had been running the plant since it was first brought on line (15+ years experience) and new operators with less than 2 years experience. More and more new operators were being added as the experienced operators transitioned out of the work force (brain drain). According to the plant manager a main goal was "knowledge transfer" within the operations staff.

Solution

exida conducted a training class on alarm management best practices to help Ashland's multifunctional team gain a common understanding and to shape their vision for the future. This was followed by a two day alarm philosophy development workshop which defined Ashland's alarm management guidelines (e.g., alarm prioritization, criteria for having an alarm, alarm classes, alarm setpoint methodology, and treatment of non-alarm notifications).

Prior to beginning alarm rationalization, the DeltaV alarm configuration was exported and loaded into SILAlarm. The CVE tool was used to analyze three years of operational data (stored in OSI PI) in order to establish actual operating envelopes and recommended alarm setpoints. This information was imported into the SILAlarm master alarm database. 

SILAlarm was then used to facilitate the alarm rationalization process and document the results. exida "kicked off" the rationalization process and trained the Ashland team so they could effectively lead the process themselves.  Key alarm rationale such as cause, consequence, corrective action, and operator time to respond were recorded for display in DeltaV Alarm Help. Rationalized alarm settings were exported from SILAlarm and imported into DeltaV.  This updated parameters such as limit, priority, class, hysteresis, on/off delay, suppression time and also configured the DeltaV Alarm Help faceplate for each alarm.

Results

The Ashland team completed rationalization of some 7000+ alarms over the course of several months (as a part-time activity).  Operators from all crews took part in the rationalization, which helped with buy-in and training. Operators saw first-hand how alarm priority would now reflect the urgency of response and the severity of the consequences.  Taking part in the rationalization process with the senior operators was educational for the novice and intermediate operators yielding significant knowledge transfer.

Key findings:

  • A large number of High-High and Low-Low alarms were eliminated since their operator response was not different from the associated High and Low alarms.
  • There was a large number alarms in the "LOPA-listed" class, which meant that numerous alarms were being used to provide risk reduction.
  • Some existing alarms were found to NOT meet the criteria for being an alarm, but were still needed by operations. These notifications were reconfigured as Alerts or Prompts so that they could be annunciated and displayed to the operator separately from alarms.
  • Displaying the results of rationalization in DeltaV Alarm Help aids novice and intermediate operators to understand the meaning and consequence of configured alarms.

Key Rationalization Results

 

Enabled Alarms

Alarms Eliminated

New Alarms

Prompts / Alerts

Alarm Priority Distribution (Critical+ / Warning / Advisory)

Before

6989

-

-

-

2.0 / 74.5 / 23.5 %

After

2163

4442

≈200

670

8.4 / 29.1 / 62.5 %