3. • Incident investigation system:
• Human errors
• Equipment failure
• Helps solve problems:
• Reactively
• Proactively
What is TapRooT®
4. TapRooT® Process
Collect Information
Understand what
happened
Identify problems
which caused the
incident
Analyze each
problem’s root cause
Search for systemic,
cultural &
organizational causal
factors
Propose corrective
actions
Present lessons learnt
to management to
obtain support for the
proposed corrective
actions
5. 1) Your root cause analysis is only as good as the information you
collect
2) Your knowledge (not only the lack of it) can get in the way of a
good root cause analysis
3) You have to understand what happened before you can
understand why it happened
4) Interviews are NOT about asking questions
TapRooT® Best Practices
6. 5) You can’t solve all human performance problems with
discipline, training, and procedures
6) Often, people can’t see effective corrective actions even if they
can find the root causes
7) All investigations are NOT equal (some investigation steps
can’t be skipped)
TapRooT® Best Practices
7. 0) Incident: safety lock applied to a water pump electrical
disconnect was cut in error by somebody other than the lock owner
Steps 1 -4:
TapRooT® Human Performance
Investigation
9. 11) Parts are ready
12) Water is shut off
13) PG supervisor leaves for the day before work is
completed. Mechanical Services (MS) group leader takes
over work coordination
15) Keys are handed over
Collect Information & Understand what Happened
10. 17) Work finished, system brought to operations conditions
18) the MS group leader gives the 2 keys to an attendant to
remove the lock
19) MS and PG supervisor communicate and PG advices MS to
cut the padlock
Collect Information & Understand what Happened
14. • Workers (PG & ES) unware of site lock-out procedure for multi-shift lock-out
• Use of personal lock for a multi-shift lock
• Padlocks hand inscribed, hard to read
• PG supervisor didn’t check which key he needs to hand to MS supervisor
• PG supervisor didn’t show MS supervisor the lock-out location
• MS supervisor didn’t tell attendant that he must remove one lock-out only
• MS supervisor didn’t tell attendant the name of the pump to be re-connected
Search for systemic, cultural & organizational causal
factors
15. • Attendant doesn’t relay that a padlock was already removed
• Attendant doesn’t ask which pump to re-connect
Search for systemic, cultural & organizational causal factors
16. a) Training Programme and policy/procedure
I. Multi-shift lock-out
II. Padlock label
b) Clear policies & procedures
c) Audits and Evaluations
d) Inforce policy on lock-out through internal audits and disciplinary actions when
applicable
e) Timely communication of policies and procedures
Propose Corrective Actions
17. • Ensure policies on Lock-out and on Device Disable are
consistent
• Group training plans to reference above policies when
appropriate
• Policy should be revised to include the acceptable methods of
marking a lockout/disable padlock
• All TRIUMF groups using shop padlocks as part of multi-shift
device disables should be made aware of current policy and
change all shop padlocks to padlocks from Operations Group
Propose Preventative Actions
18. • Although not required by WorkSafe BC, BC Safety Authority or
BC Electrical Code, it would be best practice for any electrical
wiring that has been made safe (de-energised and device disable
installed) and disconnected from a device to also have the device
end wires be covered appropriately
• Group supervisors to schedule periodic discussion sessions on
policies and procedures and invite subject matter experts able to
answer questions as needed
Propose Preventative Actions
19. • TapRooT dos and don’ts:
• Don’t assume
• Ask “what happened” not “why”
• Expect more than one root cause
• Human error is just the starting point of a root cause analysis
• Search for root causes that can be fixed
Summary
20.
21. Thank You!
Merci!
Questions?
TRIUMF: Alberta | British Columbia|
Calgary | Carleton | Guelph | Manitoba |
McGill | McMaster | Montreal | Northern
British Columbia | Queen’s | Regina |
Saint Mary’s | Simon Fraser | Toronto |
Victoria | Western | Winnipeg | York
Editor's Notes
TapRooT is a root cause analysis methodology developed and owned by System Improvements Inc.
TapRooT was created by Mark Paradies who is the president of the System Improvements Inc. Mark started as a Admiral Rickover’s Nuclear Navy officer and worked for DuPont and Westinghouse before starting the System Improvements in 1988. The first TapRooT book was published in 1991 with a focus on human performance. It analyzed the human performance to find ways to correct human factors problems - not to assign blame.
In 2000 the equipment performance investigation was integrated in the TapRooT book.
System Improvements continues to improve the system, the written material, the software and training
There are 7 steps in the TapRooT investigation process.
TapRooT promotes 7 good practices
Don’t start an investigation by guessing what happened: this will bias your questions and the outcome of the investigation. Collect all information available which can help understand the incident, the sequence of events, etc. When needed, use a SnapCharT: helps organize information collected, identify information missing.
Previous experience can bias your investigation; Looking for cause/effect
Start an investigation by asking “what happened” not “why it happen”. The purpose of the interview is to collect info not justifications for what happened. Ask the interviewee to provide the smallest and apparently least important details – they can be very valuable info for the investigation
Listen: interviews are not only about asking questions - one must listen. Don’t interrupt to ask questions: it will break the remembering process and important info might be lost
Statistically, most corrective actions for human performance problems are discipline, training and procedures. When these fail to improve performance it’s obvious that the CA were misused. The investigator should seek CAs based on the root causes which have the potential to safeguard against future human errors.
Investigator must be able to think out of the box. Doing the same thing in the same way for a long time prevents investigators to see effective corrective actions.
Root cause investigations must be tailored to the size of the problem. Simple incident investigation can be performed by one TapRooTer while more complex ones might need a team of 2, 3 or more. Some small incidents simply don’t need an investigation at all, just the implementation of a corrective action.
1) These are old type pumps which comprise a motor connected to the pump via a drive shaft.
1) Cooling system has 2 water pumps: P-1 & P-2 - none are labeled.
2) Pumps’ electrical disconnects are just behind pumps in poor lit area
3) Electrical disconnect boxes are hand inscribed with felt pen – easy to read under good lighting
4) P-1 failed and was removed for repair by Plant Group (PG). A PG lock was installed on the pump electrical disconnect
5) Decision made to buy a new pump as repair was not possible – PG lock is left in place
6) As the 2 pumps are of the same type and age decision made to replace both (buy 2 pumps)
7) New pumps are delivered
8) Electrical Services (ES) placed a lock on P-1 disconnect and removed the feed wires from the pump motor. Wires are left on the floor with bare ends exposed
9) Defective P-1 removed & new pump is installed by PG – it is found that parts need to be made (adaptor to the water line)
10) Work is halted. Two locks (PG & ES) are left on the pump electrical disconnect
Left: old water pump
Right: new water pump with integrated motor
12) PG supervisor locks the second pump (P-2) to prevent its damage by running it without water
15) PG supervisor goes to his desk to pick up the P-2 lock key: 2 keys on the desk – one for P-1 (failed) the other for P-2. PG supervisor cannot remember which is which and hands in the 2 keys to the MS group leader saying “one will unlock the padlock’
16) PG supervisor does not show the MS group leader the padlock
17) Electrical connection to P-1 is finished, system is brought to operations conditions = valves opened
18) The attendant thinks there must be 2 padlocks on the disconnect because he was given 2 keys. He removes a padlock but cannot remove the second one
22) After the loud bang the MS supervisor locks back P
The online Lockout/Tagout training did not emphasize that OG padlocks are to be used when work lasts more than one shift and the importance of labeling the padlock with the owner’s name (not only the group)
Lock-out & Disable policies confusing
Had an audit been performed, it would have shown that none of the groups were aware of the device lock-out procedures
..
Analysis of previous NCRs showed that 38% of reported nonconformities were related to policies not being used or in need of improvement. This indicates that this failure of staff to be aware of, and competent in, the application of organizational policies is not unique.