of 51
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
1/51
1
Lessons From the
Columbia DisasterSafety & Organizational
Culture
2005 American Institute of Chemical EngineersPresentation Rev_newv4_final as of 11_15_05
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
2/51
2
FEB 1, 2003 8:59 EST
All 7 astronauts are killed
$4 billion spacecraft isdestroyed
Debris scattered over2000 sq-miles of Texas
NASA grounds shuttlefleet for 2-1/2 years
Space shuttle Columbia,re-entering Earthsatmosphere at 10,000mph, disintegrates
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
3/51
3
Columbia- The Physical Cause
Insulating foam separatesfrom external tank 81seconds after lift-off
Foam strikes underside of
left wing, breachesthermal protection system(TPS) tiles
Superheated air enterswing during re-entry,
melting aluminum struts Aerodynamic stresses
destroy weakened wing
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
4/51
4
A Flawed Decision Process
Foam strike detected inlaunch videos on Day 2
Engineers requestedinspection by crew or
remote photo imageryto check for damage
Mission managersdiscounted foam strikesignificance
No actions were taken toconfirm shuttle integrity orprepare contingency plans
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
5/51
5
Seventeen Years Earlier
January 28, 1986, theshuttle Challengerexplodes 73 secondsinto its launch, killing all
seven crew members
Investigation revealsthat a solid rocketbooster (SRB) joint
failed, allowing flamesto impinge on theexternal fuel tank
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
6/51
6
Challenger
Liquid hydrogen tank explodes, ruptures liquidoxygen tank
Resulting massive explosion destroys the shuttle
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
7/517
The Legacy of Challenger
The Rogers Commission, whichinvestigated the incident, determined:
The SRB joint failed when jet flamesburned through both o-rings in the joint
NASA had long known about recurrentdamage to o-rings
Increasing levels of o-ring damage hadbeen tolerated over time
Based upon the rationale thatnothing bad has happened yet
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
8/518
The Legacy continued
The Commission also determined that:
SRB experts had expressed concerns about thesafety of the Challenger launch
NASAs culture prevented these concerns from
reaching top decision-makers Past successes had created an environment of
over-confidence within NASA
Extreme pressures to maintain launch schedulesmay have prompted flawed decision-making
The Commissions recommendations addressed annumber of organizational, communications, and safetyoversight issues
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
9/519
Columbia- The Organizational Causes
In our view, the NASA organizational
culturehad as much to do with thisaccident as the foam.
CAI B Report, Vol. 1, p. 97
NASA had received painfullessons about its culture fromthe Challenger incident
CAIB found disturbingparallels remaining at the time
of the Columbia incidentthese are the topic of thispresentation
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
10/5110
ColumbiaKey Issues
With little corroboration, management had becomeconvinced that a foam strike was not, and could notbe, a concern.
Why were serious concerns about the integrity ofthe shuttle, raised by experts within one day afterthe launch, not acted upon in the two weeks priorto return?
Why had NASA not learned from the lessons ofChallenger?
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
11/5111
1. Maintain Sense Of Vulnerability2. Combat Normalization Of Deviance3. Establish an Imperative for Safety4. Perform Valid/Timely Hazard/Risk Assessments5. Ensure Open and Frank Communications6. Learn and Advance the Culture
Key Organizational Culture FindingsWhat NASA Did Not Do
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
12/5112
Maintaining a Sense of Vulnerability
Let me assure you that, as of
yesterday afternoon, the Shuttle was
in excellent shape, there were no
major debris system problems
identified.
NASA off icial on Day 8
The Shuttle has become a matureand reliable system about as safe
as todays technology will provide.
NASA off icial in 1995
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
13/5113
Maintaining a Sense of Vulnerability
NASAs successes (Apollo program, et al) had createda can do attitude that minimized the considerationof failure
Near-misses were regarded as successes of a robustsystem rather than near-failures
No disasters had resulted from prior foam strikes,so strikes were no longer a safety-of-flight issue
Challenger parallel failure of the primary o-ring
demonstrated the adequacy of the secondary o-ringto seal the joint
A weak sense of vulnerability can lead to taking futuresuccess for granted and to taking greater risks
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
14/5114
Combating Normalization of Deviance
No debris shall emanate
from the critical zone of the
External Tank on the launch
pad or during ascent
Ground System Specif ication Book
Shuttle Design Requirements
After 113 shuttle missions,foam shedding, debrisimpacts, and TPS tiledamage came to beregarded as only a routine
maintenance concern
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
15/51
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
16/5116
Establish An Imperative for Safety
When I ask for the budget to be cut,
Im told its going to impact safety on
the Space Shuttle I think thats a
bunch of crap.
Daniel S. Goldin,
NASA Administrator, 1994
The shuttle safety organization, funded by the programs itwas to oversee, was not positioned to provideindependent safety analysis
The technical staff for both Challenger and Columbia wereput in the position of having to prove that managements
intentions were unsafe
This reversed their normal role of having to prove
mission safety
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
17/5117
Establish An Imperative for Safety
International
Space Station
deadline
19 Feb 04
Desktop screensaver at NASA
As with Challenger, futureNASA funding requiredmeeting an ambitious launchschedule
Conditions/checks, once
critical, were now waived
A significant foam strike on
a recent mission was not
resolved prior toColumbias launch
Priorities conflicted and
production won over safety
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
18/5118
Perform Valid/TimelyHazard/Risk Assessments
Anymore activity today on the tile damage or are people just relegated tocrossing their fingers and hoping for the best?
Email Exchange at NASA
hazard analysis processes are applied inconsistently across systems,
subsystems, assemblies, and components.
CAIB Report, Vol. 1, p. 188
NASA lacked consistent, structured approaches foridentifying hazards and assessing risks
Many analyses were subjective, and many action itemsfrom studies were not addressed
In lieu of proper risk assessments, many identified
concerns were simply labeled as acceptable Invalid computer modeling of the foam strike was
conducted by green analysts
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
19/51
19
Ensure Open and Frank Communications
I must emphasize (again) that severe enough
damage could present potentially grave hazardsRemember the NASA safety posters everywhere
around stating, If its not safe, say so? Yes, its that
serious.
Memo that was composed but never sent
Management adopted a uniform mindset that foamstrikes were not a concern and was not open tocontrary opinions.
The organizational culture Did not encourage bad news Encouraged 100% consensus Emphasized only chain of command communications Allowed rank and status to trump expertise
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
20/51
20
Ensure Open and Frank Communications
Lateral communications between some NASA siteswere also dysfunctional
Technical experts conducted considerableanalysis of the situation, sharing opinions withintheir own groups, but this information was notshared between organizations within NASA
As similar point was addressed by the RogersCommission on the Challenger incident
Management pushback can discourage, evenintimidate, those seeking to share concerns.
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
21/51
21
Learn and Advance the Culture
CAIB determined that NASA had not learned from thelessons of Challenger
Communications problems still existed
Experts with divergent opinions still had difficulty
getting heard
Normalization of deviance was still occurring
Schedules often still dominated over safety concerns
Hazard/risk assessments were still shallow
Abnormal events were not studied in sufficient detail,
or trended to maximize learnings
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
22/51
22
An Epilog
Shuttle Discovery was launchedon 7/26/05
NASA had formed anindependent Return To Flight(RTF) panel to monitor its
preparations 7 of the 26 RTF panel members
issued a minority report prior tothe launch
Expressing concerns aboutNASAs efforts
Questioning if Columbiaslessons had been learned
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
23/51
23
An Epilog
During launch, a large piece of foam separated from theexternal fuel tank, but fortunately did not strike theshuttle, which landed safely 14 days later
The shuttle fleet was once again grounded, pendingresolution of the problem with the external fuel tankinsulating foam
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
24/51
Turning Inward- Our Industry-
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
25/51
25
Piper Alpha
On 7/6/1988, a series ofexplosions and firesdestroyed the Piper Alphaoil platform
165 platform workers and2 emergency responderswere killed
61 workers survivedby jumping into theNorth Sea
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
26/51
26
The Physical Cause
It is believed that a pumphad been returned toservice with its dischargerelief valve removed fortesting
The light hydrocarbon(condensate) that wasreleased formed a vaporcloud and ignited
The resulting vapor cloud
explosion ruptured oilexport lines and ignitedfires on the platform
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
27/51
27
The Physical Cause
Other interconnectedplatforms continuedproduction, feeding theleaks on Piper Alpha
Ensuing fires breachedhigh pressure naturalgas inlet lines on theplatform
The enormity of the
resulting conflagrationprevented any organizedevacuation
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
28/51
28
The Organizational Causes
The official investigation report, written by LordCullen, faulted the companys management of safetyon Piper Alpha
The confusion leading to restarting the condensatepump resulted from failures to adhere to the permit towork (PTW) system
Daily monitoring and periodic audits had failed to
identify the continuing dysfunction of the system
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
29/51
29
The Organizational Causes
Inadequate shift turnovers failed to communicate
the status of the pump to the oncoming shift
Inadequate communications (and PTW systemproblems) had contributed to a fatality, and a
civil conviction for the company, but remedialaction had not been taken
The diesel fire pumps were in manual and, after the
explosion, could not be reached by staff seeking to
start them A prior audit recommendation to stop this
practice had not been implemented
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
30/51
30
The Organizational Causes
Even if fire water had been available, many delugenozzles were plugged
The company had been trying to resolve thisproblem for at least four years, but repairs were
behind schedule
One year earlier, an engineering study had concludedthat the gas risers were vulnerable and that a massivegas release could prevent successful evacuation of
the platform Management had discounted the study results
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
31/51
31
The Organizational Causes
Other problems that audits and management reviews hadfailed to identify and/or resolve included:
Emergency response training given to workers new to the
platform was cursory and often omitted. Some workers
had not been shown the location of their life boat.
Platform managers had not been trained on how to
respond to emergencies on other platforms (e.g., when to
stop production)
Evacuation and emergency shutdown drills on Piper Alpha
were not conducted according to schedule
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
32/51
32
Parallels to NASAand Columbia
Each Piper Alpha
organizational cause can bemapped to one or more ofthe NASA lessons
Maintain Sense OfVulnerability
Combat Normalization OfDeviance
Establish an Imperative forSafety
Perform Valid/Timely
Hazard/Risk Assessments Ensure Open and Frank
Communications
Learn and Advance theCulture
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
33/51
33
Flixborough
On 6/1/1974, a massivevapor cloud explosion(VCE) destroyed a UKchemical plant
Consequences:
28 employees diedand 36 were injured
Hundreds of off-siteinjuries
Approx. 1800 homesand 170 businessesdamaged
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
34/51
34
The Physical Cause
22 3344 66
2020--inchinch
bypassbypass 125 psi125 psi
1122 33
44 66
2020--inchinch
bypassbypass 125 psi125 psi
11
Approx. 30 tons of boiling cyclohexane released from
reactor system
Most likely release cause was the failure of atemporary piping modification
Installed between two reactors
Was a bypass for reactor removed for repairs
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
35/51
35
The Physical Cause
Bellows not designedfor 38-ton thrust
Design standards for
bellows ignored Inadequate pressure
test of installation
Inadequate vertical andlateral support for
jumper
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
36/51
36
The Organizational Causes
No qualified mechanicalengineer on-site
Inadequate concern withthe cause of the reactorfailure
Jumper connectionconsidered a routineplumbing job
No detailed designfor jumper
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
37/51
37
The Organizational Causes
Hurry upattitude of management
Overworked staffdid not take time toproperly analyzetheir actions
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
38/51
38
Parallels to NASAand Columbia
Each Flixboroughorganizational causecan be mapped to oneor more of the followingNASA lessons
Maintain Sense OfVulnerability
Establish anImperative for Safety
Perform Valid/TimelyHazard/RiskAssessments
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
39/51
39
Could this happen to us?
Optional: PasteCompany logo
here
Complacencydue to our superior safety performance
Normalizingour safety critical requirements
Ineffective Risk Assessmentsof our systems
Reversing the Burden of Proofwhen evaluating safetyof operations
Employees Not Speaking Freelyof their safetyconcerns
Business Pressuresat odds with safety priorities
Failure to Learnand apply learnings to improving ourculture
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
40/51
40
Title for Relevant Company Event
Use this section to briefly
summarize key aspectsof the event Do not addresses
causes here Add additional slides
if required
Paste photo related toevent in space at right, ifdesired
JPG files at 300 dpi,provide adequate
resolution If photo is not
provided, drag rightborder over toexpand this text box
Optional: PasteCompany logo
here
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
41/51
41
The Physical Cause
Briefly describe thefactors that caused theevent
Do not addressorganizationalfactors here
Add additionalslides if required
Add photo to the right,or expand the text boxas desired/needed
Optional: PasteCompany logo
here
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
42/51
42
The Organizational Causes
Describe the organizational causes of the event
Where feasible, lay a basis for parallels to the 6NASA organizational culture findings
Maintain Sense Of Vulnerability Combat Normalization Of Deviance Establish an Imperative for Safety Perform Appropriate and Timely Hazard/Risk
Assessments Ensure Open and Frank Communications Learn and Advance the Culture
Optional: PasteCompany logo
here
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
43/51
43
Parallels to NASAand Columbia
If you feel that thiswould add to theemphasis of themessage, include one ormore slides thatemphasize how yourorganizational causesrelate to the underlyingthemes from Columbia
Alternatively, youmay want to leave
this as an individualor group exercisefor the audience
Optional: PasteCompany logo
here
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
44/51
Indicators Of OrganizationalCulture Weaknesses
The following slidesprovideexamples of indicators that
your organization is
NOT Maintaining a
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
45/51
45
NOTMaintaining aSense of Vulnerability
Safety performance has been good and you do notrecall the last time you asked But what if?
You assume your safety systems are good enough
You treat critical alarms as operating indicators
You allow backlogs in preventative maintenance ofcritical equipment
Actions are not taken when trends of similar
deficiencies are identified.
NOT Preventing
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
46/51
46
NOTPreventingNormalization of Deviance
You allow operations outside established safeoperating limits without detailed risk assessment
Willful, conscious, violation of an established
procedure is tolerated without investigation, or withoutconsequences for the persons involved
Staff cannot be counted on to strictly adhere to safetypolicies and practices when supervision is not aroundto monitor compliance
You are tolerating practices or conditions that wouldhave been deemed unacceptable a year or two ago
NOT Establishing An
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
47/51
47
NOTEstablishing AnImperative for Safety
Staff monitoring safety related decisions are nottechnically qualified or sufficiently independent
Key process safety management positions have beendowngraded over time or left vacant
Recommendations for safety improvements areresisted on the grounds of cost or schedule impact
No system is in place to ensure an independent reviewof major safety-related decisions
Audits are weak, not conducted on schedule, or areregarded as negative or punitive and, therefore, areresisted
NOT Performing Valid/Timely
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
48/51
48
NOTPerforming Valid/TimelyHazard/Risk Assessments
Availability of experienced resources for hazard or riskassessments is limited
Assessments are not conducted according to schedule
Assessments are done in a perfunctory fashion, orseldom find problems
Recommendations are not meaningful and/or are notimplemented in a timely manner
Bases for rejecting risk assessment recommendations are
mostly subjective judgments or are based upon previousexperience and observation.
NOT Ensuring Open and
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
49/51
49
NOTEnsuring Open andFrank Communications
The bearer of bad news is viewed as not a teamplayer
Safety-related questioning rewarded by requiring thesuggested to prove he / she is correct
Communications get altered, with the messagesoftened, as they move up or down the managementchain
Safety-critical information is not moving laterallybetween work groups
Employees can not speak freely, to anyone else, abouttheir honest safety concerns, without fear of careerreprisals.
NOTLearning and Advancing
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
50/51
50
g gthe Culture
Recurrent problems are not investigated, trended, andresolved
Investigations reveal the same causes recurring time and
again Staff expresses concerns that standards of performance
are eroding
Concepts, once regarded as organizational values, arenow subject to expedient reconsideration
Engineering By View Graph
8/13/2019 Lessons From the Columbia Disaster Safety and Organizational Culture
51/51
Engineering By View Graph
When engineering analyses and risk assessments are condensed to fit
on a standard form or overhead slide, information is inevitably lost
the priority assigned to information can be easily misrepresented by its
placement on a chart and the language that is used.
The CAIB faulted shuttle project staff for trying tosummarize too much important information on toofew PowerPoint slides
We risk the same criticism here
This presentation introduces the concept oforganizational effectiveness and safety culture, asexemplified by the case studies presented
This is only the beginning