Aerospace Garage

Upskill / Flight Operations / Predictive safety and flight data

Predictive safety and flight data

From reactive to predictive safety: flight data monitoring, indicators that look ahead, data protection, shared data, analytics for precursors and new technologies. Part 4 of the Safety and human factors series.

By Aerospace Garage Desk · What's next · 50 minutes · 24 narrated pages

  1. Part 1 · Basics

    Safety management systems (SMS) in practice

  2. Part 2 · In practice

    Crew resource management and threat and error management

  3. Part 3 · Managing it

    Fatigue risk management

  4. Part 4 · What's next

    Predictive safety and flight dataYou are here

Safety and human factors · all 4 parts US$59 (save US$12) · See the Flight Path →

What you will be able to do

  • ✓ Explain why accident counts cannot guide one organisation's safety decisions, and how precursors and margins measured on normal flights can.
  • ✓ Compare the ICAO, EASA and FAA requirements for flight data monitoring, including helicopters and smaller aircraft, and judge whether a programme actually works, using the NTSB findings on the 2015 Akron accident.
  • ✓ Build distribution-based and barrier-based indicators and check them against changes in exposure.
  • ✓ Apply the agreements, gatekeeper role, retention rules and legal protections that keep flight data honest.
  • ✓ Evaluate shared data programmes and machine learning tools for precursors using base rates, precision and recall.
  • ✓ Plan the safety management of new technologies and a staged roadmap for predictive safety.

Course content

Session 1 · From reactive to predictive

1.1

  1. Why yesterday's numbers are not enough · free preview
  2. Precursors, margins and the shape of risk
  3. What makes an organisation predictive
  4. Worked example: what Samhan Air's numbers really say
Session 2 · Flight data monitoring programmes

1.2

  1. The rules: ICAO, EASA and the FAA
  2. Inside a flight data monitoring programme
  3. Helicopters, smaller aircraft and the gaps
  4. From the investigation file: Execuflight 1526, Akron, 2015
Session 3 · Indicators that look ahead

1.3

  1. From counting events to measuring margins
  2. Leading indicators on the barriers
  3. Indicators at State and regional level
  4. Worked example: did the new tailwind procedure work?
Session 4 · Trust: just culture and protecting the data

1.4

  1. Why protection comes first
  2. Agreements, gatekeepers and enforcement
  3. Retention, investigations and the limits of protection
  4. Worked example: a hard landing and a request for names
Session 5 · Sharing data and finding precursors

1.5

  1. Sharing beyond the airline: ASIAS and Data4Safety
  2. Machine learning for precursors
  3. Trusting what a model tells you
  4. Worked example: tuning an anomaly detector for the helicopters
Session 6 · Safety management of new technologies

1.6

  1. New technology, no history
  2. Artificial intelligence inside the safety system
  3. Watching a new aircraft type in service
  4. Worked example: Samhan Air's three-year predictive safety roadmap

Who this course is for

  • Flight data monitoring analysts and gatekeepers
  • Safety managers building indicators that look ahead rather than count accidents
  • Safety leaders evaluating shared data programmes and machine learning tools for precursors

You need: Safety management systems (SMS) in practice, plus experience with occurrence or flight data.

Upskill courses are for professional development and are not approved by any aviation authority. Learner Terms

Course content © Aerospace Garage Limited. All rights reserved.