BRENNEN

Industries / Transport and vehicles / Aviation / Helicopter services

Crew training and safety

This work keeps pilots and passengers alive: training, the data that shows how flights are really flown, and the rules and equipment for escaping a helicopter in the sea.

6

Milestones achieved, from 2002 to 2024.

1

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 2002

    Flight data read every day

    Flight data monitoring (FDM) means downloading an aircraft's recorded data after flights and scanning it for unsafe events. The UK CAA's trial with Shell Aircraft, Smiths Aerospace and Bristow Helicopters ran five Super Pumas from Aberdeen and Scatsta, gathered about 11,500 hours of data and was called by the CAA "the first application of FDM to helicopters". The UK Offshore Operators Association then committed its members to put it on all UK offshore helicopters carrying flight data recorders.

    Lessonin 2016 the CAA reported that the voluntary programs had not given the insight into accidents it expected, and pushed for better event definitions before Europe's offshore rules made FDM mandatory for large helicopters from mid-2018.

    caa.co.ukcaa.co.uk

    Credit

    UK CAA, Shell Aircraft, Smiths Aerospace, Bristow Helicopters, British Airways (analysis software)

  2. 2010

    Norway measures its risk

    SINTEF's third Helicopter Safety Study (HSS-3), a joint project of Norway's petroleum and helicopter industries, unions and authorities, found one accident and no deaths on the Norwegian shelf from 1999 to 2009, against 2.3 deaths per million passenger flight hours in 1990 to 1998. Its experts still judged the real fall in risk at about 16%, after the 45% fall estimated by the previous study, and credited new helicopter types, vibration monitoring, simulator training and better helidecks.

    sintef.no

    Credit

    SINTEF (Ivonne Andrade Herrera and colleagues)

  3. 2013

    Airspeed falls unnoticed

    On 23 August 2013 a Super Puma, G-WNSB, carrying 16 offshore workers crashed into the sea on approach to Sumburgh, Shetland, and four passengers died. The autopilot held the descent while the commander set power by hand; power was too low, the airspeed fell below 35 knots and neither pilot noticed until too late. The AAIB found the instruments were not monitored effectively and that the operator's procedures did not make the best use of the automation.

    Lessonwhen automation leaves one quantity to the pilot, someone must be assigned to watch it.

    gov.uk

    Credit

    AAIB, final report March 2016

  4. 2014

    Sea-state and seating rules

    The CAA's safety review CAP 1145, done with Norway's aviation authority and EASA, counted 25 UK offshore helicopter accidents from 1992 to 2013, seven of them fatal. From June 2014 it banned flights when the sea at the destination was above sea state 6 (a significant wave height over 6 meters) and allowed passengers only in seats next to push-out window exits unless they wore approved breathing equipment or the helicopter had side floats. From September 2014 flights were limited to the sea states each type had been shown to survive in a ditching.

    caa.co.uk

    Credit

    UK CAA, with the Norwegian CAA and EASA

  5. 2015

    Shoulders measured, air carried

    To match workers to exits, the industry body Step Change in Safety measured more than 40,000 offshore workers before 1 April 2015. Just under 3% had shoulders wider than the smallest window exit's diagonal; they are marked "extra broad", wear a checkered armband and sit next to larger exits. Since 1 January 2015 all passengers have worn a Category A emergency breathing system, one that can be put on underwater within a single breath-hold, and flight crews have worn one since April 2016. A shoulder breadth measurement is still required before offshore flights in the UK.

    caa.co.ukopito.com

    Credit

    UK CAA, Step Change in Safety, Robert Gordon University, OPITO

  6. 2024

    A full cabin escapes, timed

    Rules assume everyone can get out of a capsized helicopter within 60 seconds. For EASA, CAA International filled a capsizing cabin simulator with 13 seated test subjects plus four at the back; the last seated person was out in 51.3 seconds, and those moving from the back took 13.8 seconds more. Every subject could push out a window exit set to 204 newtons (21 kilograms) by hand from an upright seat, below the 245 newtons guidance allows, and the breathing sets kept subjects calm.

    easa.europa.eueasa.europa.eu

    Credit

    CAA International for EASA, research project HUE2, 2022 to 2024

Watch for

  1. Escape research into the rules

    After the HUE2 research, EASA told the AAIB it would review the researchers' proposals for updated guidance on exit forces and escape by reopening its rulemaking task on offshore helicopter ditching, RMT.0120, or by adding them to another helicopter rulemaking task. The AAIB rated the response "partially adequate".

    Achieved whenEASA publishes amended certification guidance for underwater exits and escape that uses the HUE2 results.

    assets.publishing.service.gov.uk

    Claimed by

    EASA, response dated November 2024 and published by the AAIB in May 2025

Open

  1. 01

    A public low-energy approach set

    A public set of time series (airspeed, height, power and rate of descent) digitized from the flight-data plots printed in accident reports where a helicopter lost energy on approach, alongside normal approaches printed in the same reports, plus an open reference alert rule and a table of how many seconds of warning it gives on each.

    The brief
    Credit · unclaimed
  2. 02

    Skill decay from training records

    An analysis of anonymized refresher-course records from offshore survival training centers: for each trainee, the months since their last helicopter underwater escape training (HUET) course and how they performed on each escape exercise on return, where the center records it, with a model of how performance falls as the gap grows.

    The brief
    Credit · unclaimed

Open 01 · the brief

A public low-energy approach set

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public set of time series (airspeed, height, power and rate of descent) digitized from the flight-data plots printed in accident reports where a helicopter lost energy on approach, alongside normal approaches printed in the same reports, plus an open reference alert rule and a table of how many seconds of warning it gives on each.

Why it is still open

G-WNSB's report prints the accident approach next to other approaches flown to the same runway, and Canada's Transportation Safety Board reports a 2019 S-92 that entered an undetected low-energy state approaching a platform. The alert envelopes for offshore terrain warning were tuned on operators' flight data, about 200,000 offshore approaches from CHC, Bristow and Babcock, which is not public, and a search found no open test set.

gov.uktsb.gc.caainonline.com

Built with what exists

WebPlotDigitizer, Python (pandas, NumPy), the example flight-data traces of normal helideck landings printed in CAA Paper 2008/02.

  1. First step

    Digitize Figures 12, 21 and 22 of the G-WNSB report.

  2. Done looks like

    At least six published approaches digitized with stated sampling and error, at least 20 normal approach traces, an open alert rule, a results table of warning time and false alerts, and 20% of the traces re-digitized by a second person with agreement reported.

Open 02 · the brief

Skill decay from training records

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

An analysis of anonymized refresher-course records from offshore survival training centers: for each trainee, the months since their last helicopter underwater escape training (HUET) course and how they performed on each escape exercise on return, where the center records it, with a model of how performance falls as the gap grows.

Why it is still open

EASA's 2020 literature review on underwater escape found "only a limited amount of research" on keeping HUET skills; Mills and Muir (1999) saw skills decay in some people at every interval from 6 to 48 months but could not find the best point to retrain, and the review repeats their call for a long-term study. Most offshore workers retrain every three to four years, an interval the CAA's 2014 review said experts widely regard as inadequate. The studies so far are small, such as a 2015 study of 38 trainees, or rest on self-assessment, such as a 2019 Dutch study of 132 air force aircrew.

easa.europa.eucaa.co.ukpubmed.ncbi.nlm.nih.gov

Built with what exists

Records held by training centers that teach OPITO's HUET standard, shared under a data agreement with names removed; Python or R for survival and logistic models; a preregistered analysis plan on the Open Science Framework.

  1. First step

    Write a one-page data request listing the fields needed and how names will be removed, and send it to two training centers.

  2. Done looks like

    Records of at least 1,000 refresher trainees from at least two centers, with performance on each exercise by six-month band since last training, a fitted decay curve with confidence intervals, and a preprint with the analysis code.