Page 17 - Index

Large Volume Data Gathering Shapes ISASI Kapustin
Aircraft Accident Investigations Scholarship Essay

By Marcin Makowski, Embry-Riddle Aeronautical University The following article is the second
of three essays from the 2021
Kapustin Scholarship winners
Aircraft accident investigation can trace its effective, have certain limitations such as that were presented during ISASI
roots to the beginning of 20th century, when the inability to analyze thousands or mil- 2021. The number of scholars
the first fatal accident involving the Wright lions of complex pieces of information that selected each year depends upon
brothers’ airplane occurred. Reactive inves- the aviation industry is capable of gathering the amount of money ISASI
tigations in those times largely focused on through the most innovative technologies members donate annually to the
scholarship fund. Details about
technological problems. that are being utilized. Safety managements scholarship applications and ad-
Later, in the mid-1960s, the focus shifted systems (SMSs) and flight operations quality ditional information can be found
more toward human factors and pilot error. assurance (FOQA) are using data analysis on the ISASI website at www.
Further on, as more and more information for threat and error management but can’t isasi.org. Application and essay
was being gathered from investigations, it comprehend all the information that could be deadlines are mid-April of each
was soon realized that many of the problems made use of. year.—Editor
could be traced back to the organizational Hence, a new model of accident investi-
level of the company involved in the accident. gation has been proposed. It consists of a
Safety investigators also realized the impor- layered structure that begins with SRBD. It
tance of predictive safety actions. [1] is raw information that was gathered in the
Within those changes, many problems system, and it is not always in usable form. It
arose, and safety investigators had to adjust needs to undergo structuring and modeling.
their techniques and develop new tools and One step higher in the model is safety infor-
models to use. The 21st century brought tech- mation. On this level, SRBD is being filtered,
nological changes that affected how all indus- and essential, useful information is extracted
tries are shaped, and this also includes the into safety information.
aviation industry. Air safety investigation is Following this process, safety law is gener-
adjusting for those changes and is constantly ated based on safety information. Safety law
challenging investigators with the need for can be defined as safety information that was
new tools, methods, and knowledge. further analyzed to predict future trends of
events. This is the proactive element of inves-
Big Data tigations. On the top of the model is safety
The modern aviation industry is driven by knowledge, which is all the actions taken
the “Internet of Things.” Multiple devices based on the previous analysis. [3]
connected with one another can interact and In the era of big data, many safety investiga-
provide users with millions of data points, tors have to be able to perform this process.
ranging from information about passengers They need to have tools to effectively gather
to airplane systems to performance, econom- data in an organized manner, clean this data,
ics, and safety. Terabytes of data are gathered organize it for investigative use, and analyze
by the airplane’s sensors, and the critical in- it. Such use of data will bring benefits in the
formation may be analyzed in near real time. form of the possibility to analyze the entire
Noncritical safety information is uploaded data set, the opportunity to analyze data from
after landing and analyzed as well. [2] Large all the resources available, and the ability to
volumes of data, arriving at the inputs with make connections between them without
significant speed, and coming from many the data being prone to interference from
different sources, are called big data. Big data the investigator’s possible natural subjective
relating to safety information is referred to as judgement. With proper tools and software,
safety-related big data (SRBD). [2] It plays a the entire process will make analysis more
fundamental role in preventive investigations efficient, more effective, less time consuming,
of aircraft accidents or incidents. and more thorough.
Accident and safety investigations focus
on the following phases: gathering of facts, Applications and Challenges
analysis of those facts, and developing con- of Using Big Data
clusions with preventive measures. In order Applications of SRBD analysis can be of
to assist safety investigators with analysis, dif- extreme use for future analysis, real-time
ferent models have been developed, for exam- analysis, and past analysis. Future and re-
ple, the Swiss Cheese Model, Human Factors al-time applications are of special interest for
Analysis and Classification System, and the the aviation industry nowadays. These can
SHELL Model. These models, while extremely include the analysis of an airplane’s perfor- Marcin Makowski
January-March 2022 ISASI Forum • 17