Get documents about "DEMONSTRATION OF A
Shared by: cuiliqing
-
Stats
- views:
- 7
- posted:
- 11/1/2011
- language:
- English
- pages:
- 30
Document Sample
LR-101-11.3-4E-94
FIELD DEMONSTRATION WORKSHOP ON
ROOT CAUSE ANALYSIS FOR
MARINE CASUALTIES AND
ENVIRONMENTAL INCIDENTS
Written by:
David A. Walker
August 2000
This report was prepared by EQE International, Inc., an ABS Group Company, for the U.S. Coast
Guard under Delivery Order Number DTCG39-99-F-E00241.
ii
SUMMARY
One of the responsibilities of U.S. Coast Guard (Coast Guard) Marine Safety Offices (MSOs) is to
conduct investigations of marine casualties and environmental incidents. Although these investigations can
have several purposes (including determining responsible parties), the investigations are an important tool
for preventing future incidents. Identifying how an incident occurred, and what can be done to prevent
recurrence, promotes continuous improvement in a waterway. Often, these investigations are loosely
structured, relying primarily on the experience and expertise of the Coast Guard investigation team to
develop findings and recommendations. Using more structured tools to provide a framework for the
investigation process may make these investigations both more effective (by more reliably identifying and
resolving the underlying root causes of incidents) and more efficient (by focusing the team’s efforts).
This report illustrates the use of two incident investigation/root cause analysis tools (namely, event
charting and root cause mapping) for more systematically performing an incident investigation. The
demonstration applications are based on an actual event involving a barge grounding and subsequent oil
spill, but the information about the event has been generalized for illustration purposes. Representatives
from the Coast Guard’s MSO Mobile, the Coast Guard’s Research and Development Center, and EQE
International, Inc. (EQE) teamed to address this topic.
The primary objective of the project was to determine whether more structured incident
investigation/root cause analysis tools would lead to more effective and efficient Coast Guard investigations
(especially when multiple stakeholders are involved). Specifically, the analysis team wanted this
application to illustrate how more structured tools could do the following:
Keep a team of diverse stakeholders (including Coast Guard, state, and local representatives)
participating in review meetings focused and productive
Integrate the many discrete, known facts (and some suppositions/questions) into one integrated
structure
Clearly identify key contributors for improvement action
Highlight where fortuitous events kept the consequences from being more severe (i.e., near misses
to a larger incident)
Identify areas where further investigation might be warranted
Uncover the underlying root causes of each significant contributor to the incident
Make recommendations (within the Coast Guard’s sphere of influence/control) for correcting the
root causes
Present the results in a format that anyone could review, understand, and/or critique
Figure 4.1 provides the event chart for the demonstration analysis as it existed at the end of the
workshop. The chart traces the sequence of events from the initiating event as the tug/barge got underway
through the grounding event, which resulted in the gasoline/oil spill. The chart continues to trace the
sequence of events from the initial oil spill through spill identification, response, and control actions
implemented by both the Coast Guard and the state Department of Environmental Protection. The event
chart identifies 10 causal factors that the team judged to be significant contributors to the incident. The
event chart also identifies one item of note revealed during the investigation.
Table 4.1 provides the summary of the root cause analysis for each causal factor and item of note. The
first column of the table lists the causal factor (or item of note) and provides a brief background description.
The second column traces the applicable path(s) through the EQE Root Cause Map™ (as determined by the
review team) and provides any comments as necessary. The third column lists suggested recommendations
for correcting the root causes (if the team made recommendations). The team believes that implementing
iii
these suggestions should reduce the frequency and/or consequences of similar incidents. However,
implementing all of the suggestions will not guarantee that such incidents will never occur. In addition,
the physical act of implementing these suggestions may create hazards that should be evaluated before
anyone tries to implement these recommendations.
The use of more structured incident investigation/root cause analysis tools helped the team to identify
some potentially important issues that were not fully explored previously. More importantly, the process
provided a structured approach for conducting the investigation and ensuring that rational risk management
strategies are in place. If state, local, and industry partners had been more actively involved in developing
the event chart and using the Root Cause Map, the analysis results could have been even better. Future
applications for marine casualties and/or environmental incidents could lead to substantial improvements in
incident rates and could foster cooperative problem solving with industry and other governmental bodies.
iv
TABLE OF CONTENTS
Section Page
SUMMARY ................................................................................................................................ iii
LIST OF TABLES AND FIGURES ......................................................................................... vii
1. INTRODUCTION ............................................................................................................. 1
2. OBJECTIVES .................................................................................................................... 5
3. APPROACH ....................................................................................................................... 7
4. RESULTS ........................................................................................................................... 9
5. OBSERVATIONS AND CONCLUSIONS ...................................................................... 19
APPENDIX A: EQE’s ROOT CAUSE MAP......................................................................... A-1
v
vi
LIST OF TABLE AND FIGURES
Table Description Page
4.1 Demonstration Analysis Root Cause Summary .......................................................... 15
Figure Description Page
1.1 Relationship Among Losses, Casual Factors, and Root Causes ................................. 3
4.1 Demonstration Analysis Event Chart ......................................................................... 11
vii
viii
1. INTRODUCTION
One of the responsibilities of U.S. Coast Guard (Coast Guard) Marine Safety Offices (MSOs) is to
conduct investigations of marine casualties and environmental incidents. Although these investigations can
have several purposes (including determining responsible parties), the investigations are an important tool
for preventing future incidents. Identifying how an incident occurred, and what can be done to prevent
recurrence, promotes continuous improvement in a waterway. Often, these investigations are loosely
structured, relying primarily on the experience and expertise of the Coast Guard investigation team to
develop findings and recommendations. Using more structured tools to provide a framework for the
investigation process may make these investigations both more effective (by more reliably identifying and
resolving the underlying root causes of incidents) and more efficient (by focusing the team’s efforts).
Typically, more structured investigation and root cause analysis tools systematically guide teams to a
more specific and defendable understanding of how an incident occurred. This understanding of how an
incident occurred is called the loss sequence and describes how specific equipment failures, human errors,
and environmental conditions (known as causal factors) contributed to the incident. While certainly
important to an investigation, causal factors are really only symptoms of underlying problems. Once the loss
sequence is known, teams use other tools to systematically search for the underlying root causes of each
causal factor. Most root causes of incidents point toward weaknesses in the management systems used to
control risks. It is generally these management system weaknesses that create (1) vulnerabilities in
equipment (e.g., inadequate preventive maintenance plans), (2) error-likely situations for people (e.g.,
confusing instructions or workplaces), or (3) vulnerabilities to extreme external conditions (e.g., capsizing
potential in certain sea states because of design flaws). This relationship among losses, causal factors, and
root causes is shown in Figure 1.1, which acknowledges that a few causal factors are simply caused by the
mistakes of fallible humans that management systems/processes cannot really prevent (at least with any
reasonable level of resources). After a team clearly identifies root causes, the team can make applicable and
effective recommendations for preventing recurrences of the incident and correcting the more systemic
management system weaknesses.
This report illustrates the use of two incident investigation/root cause analysis tools (namely, event
charting and root cause mapping) for more systematically performing an incident investigation. The
demonstration applications are based on an actual event involving a barge grounding and subsequent oil
spill, but the information about the event has been generalized for illustration purposes. Representatives
from the Coast Guard’s MSO Mobile, the Coast Guard’s Research and Development Center (R&DC), and
EQE International, Inc. (EQE) teamed to address this topic.
1
2
Problem Losses
Ca usa l Equipment Human
Fa ctors failures errors
Root Management system Personal
performance
Ca uses weaknesses issues
Figure 1.1 Relationship Among Losses, Causal Factors, and Root Causes
3
4
2. OBJECTIVES
The primary objective of the project was to determine whether more structured incident
investigation/root cause analysis tools would lead to more effective and efficient Coast Guard investigations
(especially when multiple stakeholders are involved). Specifically, the analysis team wanted this
application to illustrate how more structured tools could do the following:
Keep a team of diverse stakeholders (including Coast Guard, state, and local representatives)
participating in review meetings focused and productive
Integrate the many discrete, known facts (and some suppositions/questions) into one integrated
structure
Clearly identify key contributors for improvement action
Highlight where fortuitous events kept the consequences from being more severe (i.e., near misses
to a larger incident)
Identify areas where further investigation might be warranted
Uncover the underlying root causes of each significant contributor to the incident
Make recommendations (within the Coast Guard’s sphere of influence/control) for correcting the
root causes
Present the results in a format that anyone could review, understand, and/or critique
The ultimate objectives from the use of more structured incident investigation/root cause analysis tools
are the following:
Prevent the same or a similar incident from recurring
Prevent seemingly unrelated incidents having common underlying root causes from occurring by
fixing the root causes now
Begin coding root causes to support root cause trending to identify persistent problem areas
Promote more cost-effective and more permanent solutions by focusing on real root causes, not just
symptoms
Provide a process that (1) has a rigorous focus on factual information, (2) produces technically
defensible results, and (3) involves stakeholders in the investigation process
5
6
3. APPROACH
After participating in a loosely structured, multistakeholder review meeting that was hosted by MSO
Mobile for a specific incident, Mr. David Walker of EQE and LCDR Scott Kuhaneck from the Coast
Guard’s R&DC summarized the available data and drafted the basis for a more structured incident
investigation summary. EQE chose to use event charting to structure the causal factor identification phase
of the analysis because it has the following characteristics:
Graphically illustrates the loss sequence and highlights causal factors for further investigation
Can handle any type of contributing event (human error, equipment failure, or external
condition/natural hazard), including planned safeguards that fail
Can be effectively used in a team meeting environment using Post-It® Notes on a wall or table
Can illustrate paths that could have resulted in larger losses
EQE and the R&DC drafted the event chart through the following steps:
1. Define the loss of interest. The loss event was a spill of 1,500 gallons of gasoline and 120 gallons
of diesel fuel into a bay, which is an especially sensitive environmental area. This loss event was
described as “gasoline and diesel spill continued (1,500 gallons of gasoline and 120 gallons of
diesel),” and is shown on page two of the event chart in Figure 4.1.
2. Define the primary sequence of events leading to/from the loss event. The primary sequence
of events is (1) the series of major events that directly led to the incident and (2) the series of major
response actions taken directly after the incident. The primary event sequence is shown in bold,
horizontally across all three pages of the event chart in Figure 4.1.
3. Complete the model by adding secondary events. The various secondary event paths show the
events/conditions that caused the primary events to occur. Secondary events are added to the
structure at each step along the primary sequence of events. For each primary event, all of the
contributors leading into the event must be sufficient and necessary to cause the subsequent event
and must be supported by the facts gathered about the incident. In this way, the model of how the
incident occurred is completed. The secondary events are shown along the vertical lines on all
three pages of the event chart in Figure 4.1.
4. Identify causal factors (and items of note). The specific equipment failures, human errors, and
external conditions that significantly contribute to the incident are identified as causal factors.
Also, any special issues not contributing directly to the incident, but of concern anyway, are
identified as items of note. The causal factors and items of note are labeled in bold text on specific
boxes in Figure 4.1.
For documentation simplicity, all events and conditions are shown in Figure 4.1 as boxes.
(Traditionally, actions are shown as boxes, conditions are shown as ovals, and loss events are shown as
circles.)
After drafting the event chart using the available data, Mr. Walker of EQE reviewed this draft chart
with CAPT J. Kichner, the MSO Mobile Commanding Officer, and Mr. Bert Macesker from the Coast
Guard’s R&DC. Through the review process, a number of revisions and corrections were made to the
event chart. Figure 4.1 presents the completed event chart as it existed at the end of the demonstration
workshop. (Note that some items in this chart remain unresolved and could be explored in more detail.)
During the event chart review meeting, the EQE representative helped the Coast Guard representatives
7
use EQE’s Root Cause Map™ (provided in Annex A) to search for underlying root causes of each causal
factor. EQE chose to supplement the event charting tool with root cause mapping because it has the
following characteristics:
Systematically encourages thinking about the broad range of possible management system
weaknesses for each causal factor
Provides a structure for trending root causes across a number of incidents
Has a documentation format that links causal factors, root causes, and resulting recommendations
in an easy to understand format
For each causal factor and item of note, the EQE representative and the Coast Guard personnel
considered each branch of the EQE Root Cause Map. The team traced each applicable branch to the lowest
level of the map, which indicated a root cause that should be addressed. More than one applicable path
through the map is possible for causal factors, which indicates multiple underlying root causes. The team
used available data (and a little intuition) to determine the applicability of various paths through the map.
However, in some cases, sufficient data were not available to make a root cause determination.
For each root cause, the team discussed possible recommendations for correcting the identified
problem. In some cases, no recommendations were made because (1) no cost-effective solution was
identified or (2) any improvement action was beyond the Coast Guard’s influence/control. The EQE
representative documented the root cause analysis in the tabular format shown in Table 4.1.
8
4. RESULTS
Figure 4.1 provides the event chart for the demonstration analysis as it existed at the end of the
workshop. The chart traces the sequence of events from the initiating event as the tug/barge (T/B) got
underway through the grounding event, which resulted in the gasoline/oil spill. The chart continues to trace
the sequence of events from the initial oil spill through spill identification, response, and control actions
implemented by both the Coast Guard and the state Department of Environmental Protection (DEP). The
event chart identifies 10 causal factors that the team judged to be significant contributors to the incident.
The event chart also identifies one item of note revealed during the investigation.
Table 4.1 provides the summary of the root cause analysis for each causal factor and item of note. The
first column of the table lists the causal factor (or item of note) and provides a brief background description.
The second column traces the applicable path(s) through the EQE Root Cause Map (as determined by the
review team) and provides any comments as necessary. The third column lists suggested recommendations
for correcting the root causes (if the team made recommendations). The team believes that implementing
these suggestions should reduce the frequency and/or consequences of similar incidents. However,
implementing all of the suggestions will not guarantee that such incidents will never occur. In addition,
the physical act of implementing these suggestions may create hazards that should be evaluated before
anyone tries to implement these recommendations.
9
10
Page 1 of 3
CF1 CF2
* * * *
Channel clearance Did the tug Was the T/B Was the tug
Was some other
was reduced by experience overloaded or captain T/B had a
piloting mistake
shoaling in a few mechanical improperly distracted or single skin
made?
places difficulty? loaded? asleep?
T/B's #1 starboard
tank (gasoline)
T/B got T/B approached T/B struck Gasoline and
underway the bridge ground
and #3 starboard
diesel spill began A
Note: Tank damage may have tank (diesel)
occurred at initial ground contact damaged
or the subsequent grounding
T/B forced T/B ran aground
T/B tried to avoid
off course by before hitting the
hitting the bridge
initial grounding bridge
T/B could have hit
the bridge
* Unresolved questions that
could be explored further
The bridge could
T/B's tanks could
have been
have ruptured,
damaged,
resulting in a
resulting in
CF refers to a "causal factor," which is a specific major diesel and
property loss or
controllable equipment malfunction, human mistake, gasoline spill
traffic problems
or other condition that allowed the incident sequence
to progress.
Note: Dashed lines indicate how the
incident could have credibly escalated
into a more severe event
Figure 4.1 Demonstration Analysis Event Chart
11
Page 2 of 3
CF6 CF7
*
USCG experienced
internal What led to
USCG aircraft was
communication the aircraft
problems between unavailable on
unavailability
MSO and air demand
on demand?
station personnel
CF8
USCG chose a
shallow area to
deploy equipment
CF9 CF10
Internal DEP
DEP did not
T/B provided USCG air asset USCG water communication
receive accurate
inaccurate deployed less assets deployed problems kept
spill information
information to rapidly than less rapidly than parties from
from the marine
the USCG desired desired receiving accurate
patrol
spill information
USCG responded DEP responded
Spill occurred at a
less rapidly than less rapidly
remote location
desired than desired
CF3 CF4 CF5
Spill could have USCG response
team and T/B crew Decreasing head
USCG/DEP been larger if
T/B delayed Marine patrol did transferred in tanks slowed
T/B did not detect responded less barge did not have
reporting while not detect the gasoline and diesel and eventually
the releases rapidly than excess capacity, to empty tanks on
trying to free itself releases stopped the
desired which is really board, reducing the releases
unusual volume of release
(Identification issues) Gasoline and (Response issues)
USCG response
diesel spill
team detected the
continued (1,500 Gasoline and
A gallons of gasoline
releases upon
diesel spill stopped B
inspection at the
and 120 gallons of
scene
diesel)
Spill could have
continued even Spill could have
longer if USCG been much larger
had not chosen to if the release had
investigate not been stopped
Figure 4.1 Demonstration Analysis Event Chart (cont’d)
12
Page 3 of 3
T/B moved to T/B waited for tide T/B continued
T/B transited
deeper water for T/B tanks patched change at transit to offloading
B T/B refloated
underwater temporarily
toward offloading
entrance to location without
location
inspection offloading location incident
T/B transit path T/B delay for tide
could have endan- increased exposure
gered sensitive to a subsequent
spill, but waiting
areas if a further until daylight would
spill had occurred have further
during transit increased exposure
ION
County officials were
not kept informed nor
were their extensive
communications
systems utilized
ION refers to an "item of note," which did not directly lead to the losses
experienced with this event, but could contribute to future problems if
not addressed.
Figure 4.1 Demonstration Analysis Event Chart (cont’d)
13
14
Table 4.1 Demonstration Analysis Root Cause Summary
Root Causes
Causal Factors (Categorized Using the EQE Root Cause Map) Recommendations
CF 1: Channel clearance was reduced by shoaling in a Other difficulty None
few places Natural phenomena
Background:
Shoaling naturally occurs at various locations. The Corp
of Engineers manages this as well as possible.
CF 2: T/B had a single skin Equipment difficulty None (phase out of single-skin barges
Equipment design problem is already underway)
Background: Design input less than adequate
Many older barges have single skins that are more
vulnerable to spills from groundings, collisions, and
allisions. Because of new regulations, operators are
phasing out the use of single-skin barges.
CF 3: T/B delayed reporting while trying to free Personnel difficulty Highlight the delayed reporting to the
itself Barge company employee crew and the operating company for
Personal performance internal corrective action.
Background:
The barge crew did not report the grounding until after “Rewards/incentives less than adequate” and “problem Share the Coast Guard’s concern
they could not free the barge. The grounding should reporting less than adequate” under the about this issue with other local barge
have been reported earlier, but human nature is not to “administrative/management systems” branch were operators and encourage them to
report such an event unless there is a real problem. considered, but the problem seemed to be specific to the emphasize timely reporting to their
crew, not the existing systems. crews.
15
Table 4.1 Demonstration Analysis Root Cause Summary (cont’d)
Root Causes
Causal Factors (Categorized Using the EQE Root Cause Map) Recommendations
CF 4: T/B did not detect the release Personnel difficulty Have the barge company improve its
Barge company employee guidelines (and associated training)
Background: Administrative/management systems for assessing damage after a
The barge crew did not conduct a successful Problem identification/control grounding, collision, or allision.
investigation to determine whether oil was leaking from Problem analysis less than adequate
the barge. The crew did look for a release, but their Share the Coast Guard’s concern
investigation was not effective. Of course, it was dark, “Training difficulties” and “personal performance” were about this issue with other local barge
and a small release rate would be somewhat difficult to also considered here, but data were not available to operators and encourage them to
detect (especially if the T/B were still trying to free assign other root causes. Further investigation of this evaluate their own guidelines for
itself). The crew’s initial report was “all intact, no causal factor could be beneficial. similar weaknesses.
injury, no spill, and no bridge impact.” Personnel difficulty None (additional lighting requirements
Barge company employee would not really be practical)
Human factors engineering
Work environment
Ambient conditions less than adequate (dark)
CF 5: Marine patrol did not detect the releases Personnel difficulty Offer to provide additional training on
Marine patrol employee spill identification to the marine
Background: Administrative/management systems patrol.
The marine patrol that initially investigated the Problem identification/control
grounding (after it was reported) did not detect the Problem analysis less than adequate
releases.
“Training difficulties” and “personal performance” were
also considered here, but data were not available to
assign other root causes. Further investigation of this
causal factor could be beneficial.
Personnel difficulty None (additional lighting requirements
Marine patrol employee would not really be practical)
Human factors engineering
Work environment
Ambient conditions less than adequate (dark)
16
Table 4.1 Demonstration Analysis Root Cause Summary (cont’d)
Root Causes
Causal Factors (Categorized Using the EQE Root Cause Map) Recommendations
CF 6: USCG experienced internal communication Personnel difficulty Improve training in cross-
problems between MSO and air station personnel USCG staff (company employee) programmatic issues for “O”
Communication personnel (e.g., oil spill response plans
Background: Unknown and priorities).
The requested air asset (25B) out of Corpus Christi was
delayed because of communication issues between the Further resolution of the communication difficulty was
MSO and the air station. The air station questioned the not possible based on the data available, although a near
MSO’s request for the asset and may have had root cause of “misunderstood communication” and a
conflicting priorities to address. It is also possible that root cause of “long message” may be applicable. Further
the original request may have been misunderstood. investigation of this causal factor could be beneficial.
CF 7: USCG aircraft was unavailable on demand Unknown None at this time
Background: Further investigation of this causal factor could be
The H-65 requested by the MSO was not available for beneficial, but is somewhat outside of the
service. The cause of the aircraft’s unavailability is not control/influence of the MSO.
known to the MSO.
CF 8: USCG chose a shallow area to deploy Personnel difficulty Update the area response plan to
equipment USCG staff include accurate information about the
Procedures site.
Background: Wrong/incomplete
The USCG could not launch land-based small boats Facts wrong/requirements not correct Check other possible deployment sites
from the selected deployment site in the field because mentioned in the response plans to
the water was too shallow. The information about the ensure that similar problems do not
deployment site in the area response plan was exist for other sites.
inaccurate.
CF 9: DEP did not receive accurate spill information Unknown None at this time
from the marine patrol
Further investigation of this causal factor could be
Background: beneficial, but would have to be conducted by DEP.
Unknown
17
Table 4.1 Demonstration Analysis Root Cause Summary (cont’d)
Root Causes
Causal Factors (Categorized Using the EQE Root Cause Map) Recommendations
CF 10: Internal DEP communications problem kept Unknown None at this time
parties from receiving accurate spill information
Further investigation of this causal factor could be
Background: beneficial, but would have to be conducted by DEP.
Unknown
ION: County officials were not kept informed nor Personnel difficulty Incorporate the use of countywide
were their extensive communications systems utilized USCG/DEP staff communication systems into area
Communications response plans.
Background: No communication or not timely
The county had extensive communications systems that Communications between work groups less than
were not utilized during the event and could have adequate
improved the overall response to the incident.
18
5. OBSERVATIONS AND CONCLUSIONS
The use of more structured incident investigation/root cause analysis tools helped the team to identify
some potentially important issues that were not fully explored previously. More importantly, the process
provided a structured approach for conducting the investigation and ensuring that rational risk management
strategies are in place. If state, local, and industry partners had been more actively involved in developing
the event chart and using the Root Cause Map, the analysis results could have been even better. Future
applications for marine casualties and/or environmental incidents could lead to substantial improvements in
incident rates and could foster cooperative problem solving with industry and other governmental bodies.
19
20
ANNEX A
EQE’s Root Cause Map™
A-2
