Transcript Student Video -Aviation Safety Metrics (Bati)

Firdu Bati
Video Presentation on
Aviation Safety
Introduction
• Aviation safety is very important to the viability of
the industry in general
• Frequent Fatal accidents can have high negative
effect in the perception of the flying public and
damages the industry’s reputation
• Fortunately, aviation accidents are very rare, and
only a fraction of those accidents result in
fatalities
• The likelihood of being killed in aviation accident
is extremely small, in the order of millionth
Death Risk
• An objective measure of aviation safety is the death risk of
flying
• How likely it is a passenger wouldn't survive any randomly
chosen flight
• To determine the death risk, we rely on statistics of various
kind.
• We need to be careful about the type of statistics to
consider for death risk analysis
• Static of “fatal accidents per hours flown” may not be a
good measure.
- The variable “fatal accident” may represent an accident
with a single fatality or entire passengers of the flight with
the same weight.
- Similarly, “hours flown” ignores some important facts such
as most accidents happen in certain phase of flights
regardless of number of hours flown
Death Risk
• “Hull loss” statistic is more useful to aircraft manufactures
than assessing death risk
- An accident with heavy damage to the aircraft might
involve no deaths at all
• “Passengers killed” to “passengers carried” statistic is a
good measure of proportions, but still don’t truly measure
death risk
- The risk is not proportional by the number of passenger in
a given accident, i.e. accident with higher number of
fatalities is not proportionally riskier than with small
number of fatalities
• In general, it is better to avoid directly involving the
number of deaths in the numerator of death risk statistics
Death Risk Using Q-statistic
• The Q-statistic answers the probability of death given
a random flight of N flights with xi fraction of fatalities
- Let there be N flights to choose from
- A flight i is picked with 1/N probability
- xi is the fraction of deaths in flight i
- The conditional probability of death in flight i is xi /N
- Full Loss Equivalent-FLE is the sum of the fraction, ∑xi
- Overall probability, which is the death rate over all N
flights is given by:
Q = ∑i=1-N xi /N = FLE/N
Advantages of Using Q-statistic
• Q-statistic allows to treat accidents with
higher survival rate differently than with lower
survival rate
• It doesn’t give different weights to Distance
flown and duration of flight
• The conditional probability calculation
involved in Q-statistic is very easy
• Q-statistic calculated using recent data can be
used to estimate death of on other flights
Calculated Q-values
• Using data of scheduled passenger jet flights
in First World countries
– In year 1990 to 1999, 70 million flights performed
with only 10 of them involved in fatal accidents
– Average fraction of the 10 flights is 56%
– Q= 10*0.56/70,000,000 = .00000008 which
equates to approximately 1 fatality in 13 million
flights, a very, very small fraction
– With such a small rate, a person can fly everyday
for 36,000 years before killed in a flight accident
Calculated Q-values
• Recent statistics show even impressive results
- From 2000 to 2008, there were only 2 accidents,
one in Kentucky in 2006 with 100% fatality and
another one in Madrid in 2008 with 90% fatality
of 80,000,000 flights total
- Average fraction of the 2 fatal flights is 95%
– Q= 2*.95/80,000,000 = .000000024 which equates to
approximately 1 fatality in 42 million flights
– With such tiny rate, a person can fly everyday for
more than 116 thousand years before killed in a flight
accident
Note: Not considering survivability, the death risk tends
to be higher
Statistic Across the World
• First World Domestic
- 1960-1969
1 in a million
- 1980 – 1989
1 in four million
- 2000 – 2007
1 in 80 million
• First World International
- 1960-1969
1 in 200 thousand
- 1980 – 1989
1 in four million
- 2000 – 2007
1 in 9 million
• Between First and Developing World
- 1960 – 1969
1 in 200 thousand
- 1980 – 1989
1 in 600 thousand
- 2000 – 2007
1 in 1.5 million
• Within Developing World
- 1960 – 1969
- 1980 – 1989
- 2000 – 2007
1 in 100 thousand
1 in 400 thousand
1 in 2 million
Comparing Safety Among Operators
• With such a small overall death rate in the First World, it would be
difficult to make detailed comparison among individual airlines
• One approach would be to compare the rate of precursors to
accidents
• Using Incidents and non-fatal accidents to compare death rate
among airlines however might not reflect the true picture.
• Typically, airlines learn from incidents and take corrective actions
to minimize the occurrence of accidents from same operational
errors
• Even when a First World airline is compared with Third World
airline in the same environment/route, the rate is surprisingly
similar.
• Hence, there is no statistically significant data to claim that one
airline is better than another in terms of safety
Future Aviation Safety Assessment
• The fact that aviation accidents rate have reduced to such level in
recent decades is a testament to the safety record of the industry
• However, how do we improve safety further in the presence of
almost non existent accidents from which operators and regulators
learn and take actions to prevent such happenings
• Because of that, since the early 1970 the industry and regulators
started compiling data about incidents which are precursors to
accidents
• There are over 70 safety action programs run by airlines, NASA and
the FAA to collect data that lead to incidents which have the
potential to lead to accidents
• As a result there are enormous amount of data to analyze and
assess the future of aviation safety, and the likelihood of
catastrophic accidents
• Linear and Quadratic models can be used to assess runway and mid
air collision risks of future operations
Conclusion
• The statistics undeniably tells the whole picture, safety of
aviation has been improving at an impressive rate in spite
of increasing service
• However, the system is not completely free of accidents,
when aviation accidents do happen, they leave large dents
in the industry and erode public confidence
• We have to continue to improve the system as challenging
as that maybe
• With the expected increase of aviation services and the
introduction of new procedures and application of latest
technologies, the challenges will be even greater
• No matter how challenging the task is the safety of aviation
should not be comprised and needs to continue to improve
to sustain the viability of this vital industry
References
Chapter 11, Aviation Safety from Global
Airlines Industry by Belobaba, Odoni, and
Barnhart
Aviation Safety Lecture Notes by Dr. Lance
Sherry