Results before and after dynamic alarm management

1. Results Before and After
Dynamic Alarm Management
Dustin Beebe –ProSys
Bob Wingard – Huntsman Corporation
Bob Townley-Smith – Huntsman Corporation

2. Presenters
 Dustin Beebe
 Bob Wingard
 Bob Townley - Smith

3. Huntsman Performance Products

4. Huntsman Performance Products
Morpholine
DGA
E7
Ethylene
Oxide
F4 & F6
Ethanol-
Amines
E4
Ethylene
Glycols
G1/3/5/6
Propylene
Glycols
G2
Surfactants
E3 & E6
Olefins
A3
PO/
MTBE
F5
Oxygen
Ethylene
Ethane/Propane
FCCU Gas
Refinery Grade
Propylene
Oxygen
Isobutane
Methanol
EO
Monoethanol Amine (MEA)
Diethanol Amine (DEA)
Triethanol Amine (TEA)
Triethanol Amine Residue
Ethylene
Water
MEG
TEG
DEG
Sales
Propylene Glycol (PG)
DiPropylene Glycol (DPG)
Morpholine
EO
Water
PO
EO
NH3
Nonyl Phenol Ethoxylates
Sodium Isethionate
Amine Ethoxylates
AD
C4’s
Propylene
PO
MTBE Alcohols
Amines
MNP
Shipments Biodiesel
Normal C4
AFG
Linear Alcohol Ethoxylates
Shipments
NH3
Raw Materials
Purchases
Pipeline Barge
Rail Trucks
SO2
Propylene
EO

5. Huntsman Performance Products
Units Products Capacity Technology
A3 Ethylene 450 MMPPY 65/Stone & Webster
A3 Propylene 300 MMPPY 65/Stone & Webster
F4, F6 Ethylene Oxide 1000 MMPPY 63/68 Scientific Design
G1/3/5/6 Ethylene Glycol 890 MMPPY 47-89/In House
G2 Propylene Glycol 144 MMPPY 58/In House
E4 Ethanolamines 400 MMPPY 79/07 In House
E7 Specialty Amines 50 MMPPY 67/In House
E3/E6 Surfactants 470 MMPPY 52-71/In House
F5 PO 566 MMPPY 94/In House
F5 MTBE 1782 MMPPY 94/In House

6. Port Neches Control Systems
 Approximately 10,000 analog IO
 3000 Controllers
 73 Control Processors
 More than 35,000 blocks with at least one alarm
configured. (Most have several alarms configured.)
 36 PLC processors.
 Control is split between DeltaV and Foxboro IA

7. We have met the enemy and he is us.
Pogo
Oliver Hazard
Perry

8. Common Mistakes / Proof that the enemy is was us.
 Every SIS action was a critical alarm.
– Removed alarms from individual trips.
– One common low priority SD alarm per system.
• Operator action to check first out display.
– Save higher priority alarms for SIF failures.
 Every shutdown had a pre-alarm that was not suppressed
when the associated trip is active.
– Delay pre-alarms and suppress when associated trip is active.
 Almost every temperature on distillation column had an
alarm and the pressure.
– Choose single most critical temperature and the overhead
pressure.

9. Common Mistakes / Proof that the enemy is was us. (cont)
 Alarms were configured on individual pieces of
redundant equipment.
– Example: 6 redundant Fireye flame detectors generate 18
alarms each time the incinerator is shut down. One for each
detector and one that indicated that both detectors agree or
disagree that the fire is out.
• Remove alarms from individual Fireyes, suppress flame out alarm
when gas valve is blocked in. Create flame detected alarm when
gas valve is blocked in.
 Low Flow alarms set at less than 10% of span.
– Still have some of these but dynamic management reduces
chatter.

10. Alarm Philosophy
 Philosophy development should be your first step on your road to
alarm management.
– There are a lot of good resources on what should be in your alarm
philosophy, make use of them.
 Spend time on it and generate a good document.
– You are not helping yourself if your philosophy is simply a checklist of the
ISA 18.2 sections.
– Specific to your site.
 Will guide your teams through the rationalization process.
 Defines the life cycle processes use to maintain your alarm
systems.
 Contains clear examples that your teams can refer to during
rationalization.
 Very important that your philosophy formalize alarm management
responsibilities within the operating units.

11. Alarm Management Champions
 One for site and one for each operating unit.
 Creditability within the operating units.
 Leverage champions creditability to help reshape
wrong ideas about alarms.
 Train first, force multiplier.

12. Training
 Train your teams on your Philosophy, before you
start. Focus on understanding of the definition of an
alarm and team members roles and responsibilities.
 Ensure that each team member understands his/her
role in alarm management.
 Time spent on training up front will come back to you
during your rationalizations.

13. Implementation
 Do not put anything in front of your operators, until you are sure
it is correct.
– Spend extra time up front verifying your implementation, especially on
your first couple of units.
 If you lose creditability with you console operators, regaining
that creditability will be harder the second time.
 Bad news travels much faster than good.
 Don’t tie your operators hands, initially.
– At Huntsman we initially allowed our operators to continue suppressing
alarms the old way in parallel with the new way. Once we are sure
everything works as designed we remove the “old” ability.
 It will be difficult to meet the ISA requirements without some
form of dynamic management of your alarm settings.

14. Ongoing maintenance of your Alarm
Systems
 To be successful you must have accountability for alarm
maintenance functions.
 Management of alarms should become just another
expected job function for the personnel assigned to your
units.
– At Huntsman our alarm performance reports are part of each
units weekly KPI meetings.
 Beware of your PHA/LOPA and Project teams, ensure that
they are trained on your alarm philosophy. One bad
LOPA/Project can eliminate your gains.

15. Introduction
 Define an Alarm Flood
 What Causes Floods?
 Impact of Alarm Floods on your plant
 Dynamic Alarm Management and its ability to
control alarm floods
 Results before and after Dynamic Alarm
Management

16. Definition of Alarm Flood
Alarm Flood defined by ISA 18.2 as -
“10 or more annunciated alarms in any 10
minute period per operator”

17. What causes a Flood?
Alarms are typically configured for run
therefore many alarms are triggered
upon a Change of Process State:
• Run to Shutdown
• Plant State 1 to Plant State 2

18. Impact of Alarm Floods
 Plant Managers should be ready for alarm
performance to be scrutinized during OSHA audits
 Reports show that 70% of plant incidents occur on
startup, shutdown or transitions.
 Could incidents be caused by critical alarms being
hidden under a flood? What about before alarms have
cleared and operators are starting up the plant
without alarms?
 Even your SIS alarms are not immune to influence by
alarm floods. Even special sounding alarms can be
missed when hundreds of alarms are sounding.

19. Impact of Alarm Floods
 What about product quality, plant profitability and
equipment damage – have any of these suffered when
alarm floods were a significant distraction for the
operator?
 Have you performed an incident review to find that a
critical alarm was missed? Was a flood of alarms even
considered as a distraction for the operator?
 Do you have any data on how many alarms are
missed due to distractions?
 How many loss of containment incidents, injuries or
worse can be tracked back to an alarm flood?

20. ANSI / ISA 18.2 Alarm Metrics
Metric Target Values
Average Annunciated Alarms per Hour 6
Average Annunciated Alarms per 10
minutes
1
Percentage of Hours containing more
than 30 alarms
<1%
Percentage of 10 Minute periods
containing more than 10 alarms
<1%
Maximum number of alarms in 10
minute period
<=10
Percentage contribution of the top 10
most frequent alarms to the overall
alarm load
<1% to 5% maximum, with action plans to address
deficiencies
Percentage of time the alarm system is
in flood condition
<1%
Quantity of Chattering and fleeting
alarms
Zero, with action plans to correct any that occur
Stale alarms < 5 present on any day, with action plans to address
Annunciated priority distribution
3 priorities - 80% Low, 15% Med, 5% High, or
4 priorities - 80% Low, 15% Med, 5% High, <1% Highest

21. Dynamic Alarm Management
 Improved Quality of Alarms
– Every alarm should:
• Be clear and relevant to the operator
• Indicate an abnormal process condition that has consequences of inaction
and a defined response
• Be unique
 Correct Technology, Methodology and Resources
– Plant is divided into systems and sub-systems
– Complete and thorough review of:
• Plant operating procedures
• P&ID’s
• PHA documentation
• Interviews with engineers and operators
– Quality of process rises and falls with the quality of the
rationalization facilitator
• Experienced process engineer with P.E.
• Alarm management experience
• Capable of challenging participants to keep process on track

22. Dynamic Alarm Management
 Dynamic rationalization is rationalization for more than one
process state
 Static rationalizations can become dynamic when the question
“When” is added to the discussion for each point
 Control of alarm floods is vastly more important than improving
average alarm rates
 Dynamic Alarm Management is the only way to control alarm
floods
 Dynamic Alarm Management
– Transition manager configurable for every alarm
– Smooth transitions from state to state
– True Dynamic Alarm Management software able to handle 700+ dynamic
points per operator
– No problem meeting or exceeding ISA 18.2 metrics

24. Online Solution

28. Results Before Dynamic Alarm Management
 Peak Alarm Rate – 40 per 10 min
 Average Alarm Rate – 1.4 per 10 min
 % of Time in Flood Condition – 6.45%
 Chattering Alarms
– 78.3% of all alarms
– G5-FC1505.PVHI – 85.9% of count and 88.9% of duration

29. Results Before Dynamic Alarm Management

30. Results Before Dynamic Alarm Management

31. Results After Dynamic Alarm Management

32. Results Before Dynamic Alarm Management

33. Results After Dynamic Alarm Management

34. Results After Dynamic Alarm Management

35. Results After Dynamic Alarm Management

36. Summary
 Alarm System Design Changes
Item Before After % Diff
Alarms Configured 3745 1112 -70%
Log 0 767 +∞
Low 2003 917 -54%
Medium 1648 163 -90%
High 94 32 -66%

37. Summary
 Alarm System Design Changes
Item Metric
Trip Point Changes 357
Priority Changes 970
Alarms Eliminated 2633
Dynamically
Managed
143
(13%)
Dynamic Actions 510

38. Summary
 Performance versus ISA 18.2 Metrics
Metric Before After Targets
Avg. Alarms/10 min >24 0.2 1
Peak alarms/10 min >200 18 10
% Time in Flood 19% 0.11% <1%
Stale Alarms 72.2 7 <5
% Hours > 30 alarms 18.4% 0.1% <1%

39. Results After Dynamic Alarm Management
 Dramatic absence of redundant and unnecessary
alarms
 Operators expressed concern but when questioned -
were receiving all legit alarms
 No issues as a result of fewer alarms

40. Business Results Achieved
 Lower risk with fewer floods
 Lower operator loading
 Better operator focus

41. Summary
 Definition of alarm flood and causes
 Impact of alarm floods on your plant
 Alarm flood results before and after dynamic
alarm management
 Compliance with ISA 18.2 as a result

42. Where To Get More Information
 ProSys Contact Info
sales@prosys.com
prosys.com
225-291-9591 x225