Archive | April, 2019

The (un)Lucky 13 Aviation Safety Commandments!

13 Apr

The 13 (un)Lucky Commandments of Aviation Safety

Ethiopian Airlines Flight Remains post-Crash, March 2019

Fellow Aviation Engineers – I was discussing recent aviation headlines with our AFuzion senior engineering staff and it’s unanimous: the past six months are the historic Tipping Point which will be spoken of for the next 100 years. Remember your school Physics class and the video of the failing Tacoma Narrows Bridge emphasizing the need for safety via resonant frequencies? If you don’t know that one, you are either not an engineer or you skipped class that day.  Similarly, most of us have heard of the 10 Commandments received by Moses … Well, did you know there are ‘AFuzion’s Thirteen Commandments of Aviation Safety 2020”? If not, keep reading …

Yes, times were simpler back in Moses’ day – the aviation technology world is far more complex so an (un)lucky 13 Commandments are required. Our aviation great grandchildren will be discussing 2019 one hundred years from now as well. Perhaps we will have actually learned from these Thirteen Commandments of Aviation Safety for 2020. Yes, these are directed at the Boeing 737 MAX disasters, and specifically how we can learn from and apply these Thirteen Commandments; just as bridge design changed after the Tacoma Narrows disaster, aircraft safety, design, and oversight will be forever changed starting today. Hopefully. But only if we all understand and apply these Commandments. So here, I’ve summarized them for all current and future aviation engineers: for the next 100 years …

The (un)Lucky 13 Aviation Safety Commandments:

1.      Follow mandatory ARP4761 for Development Assurance Level (DAL) Assignment. When the safety assessment provided to FAA and EASA states the 737 MAX MCAS will only adjust the Horizontal Stabilizer by 0.6 degrees and then the airframer changes it during flight testing to 2.5 degrees x 2 activations = 5 degrees, that is a fundamentally different design and Major safety impact. Such requires a formal Functional Hazard Assessment (FHA) update and a DAL reassessment, in this case resulting in a DAL of “A” (1 x 10—9) probability instead of DAL B (1 x 10—7) or DAL C (1 x 10—5) as was originally specified and approved. Remember: the United States 14 Code of Federal Regulations (CFR) 25.1309 is not optional and this failure violates the “Update Safety Assessment Continually Commandment”.

2.     Utilize mandatory redundancy for critical systems. When people can die, the Development Assurance Level is A if the failure can cause a plane to crash or Level B if passengers can die. For Level A and B, the required reliability is met via redundancy. Yes, the 737 MAX was equipped with two sensors (redundant), but MCAS only actively used one of the sensors thus violating the “Redundancy Commandment”.

3.      Follow mandatory ARP4754A processes for System Requirements. Software is only as good as the System requirements mandated to precede it. Those System requirements must include Safety and derived requirements from the FHA and Preliminary System Safety Assessment (PSSA) per ARP4754A. This includes power-up testing of the Angle-of-Attack sensor, redundancy requirements, and pilot display/awareness requirements. The 737 MAX system requirements missed this Commandment therefore its software was doomed; subsequent patching (changing) of software to address missing Systems requirements is just another form of broken “System Commandment”.

4.     Implement both Continuous and Power-up Built-In-Test (BIT) on Power-up to test both Angle-of-Attack (AOA) Sensors and mismatches. When the sensor angle outputs differ by more than a nominal calibration amount (1-2 degrees), either both sensors should be deactivated with an accompanying pilot annunciation and MCAS deactivation, or a 3-sensor voting design should instead be deployed as befitting the MCAS system’s true DAL A FHA designation. Failure to consider these is a clear violation of “ARP4754A’s Continuous Safety Update Commandment”.

5.      Follow mandatory System Safety Assessment (SSA) ARP4754A/4761 FMEA and MTBF calculations to determine the AOA sensor reliability was insufficient for DAL B let alone DAL A; use this finding to update the design and apply to an updated PSSA thus requiring active sensor redundancy usage. Failure to do this violates the “ARP4761 FMEA/MTBF Validation Commandment”.

6.      Implement Built-In-Test (BIT) on Power-up to test both Angle-of-Attack (AOA) Sensors and mismatches. Evidence indicates the active LionAir sensor was off by 20-degrees while the plane was still taxiing yet such result was not actively used in MCAS system deactivation or pilot annunciation; this violates the “ARP4754A Derived Safety Requirement Commandment”.

7.     Follow rules for mandatory display of aircraft safety conditions including explicit pilot annunciation of MCAS Activation. When flight data recorders yield soon-to-be-dead pilots thumbing through operating manuals to determine what’s going on, it’s clear the 737 MAX pilots were not informed of MCAS system activation whose AOA sensor failure plunged them to the surface. Boeing previously had an enviable record deploying aircraft where the pilot exercised primary control versus a more automated-avionics approach. While either approach is feasible, the pilot-in-command strategy (historic Boeing and 737 MAX training protocols) require the 737MAX pilots to be informed of MCAS activation. They were not and the resultant crashes result from violating the “Promote Pilot Situational Awareness Commandment”.

8.     Follow New Aircraft Cert rules for New Aircraft. When the 737 MAX aircraft revision and heavy new engines so significantly changes the Center of Gravity thus requiring a new system (MCAS) to compensate, this is a Major design change. Adding an all-new system to mitigate the 737 MAX’s higher stall probability is good, but this yields a New design with mandatory higher recertification standards and requisite mandatory pilot retraining with significant procedural and operating manual updates. Sneaking in a Major change via a Minor update violates the “Certification Transparency Commandment”.

9.     Pilot workload and situational awareness are the single leading cause of aviation safety incidents; in my personal pilot ground-school training, pilot procedures and situational awareness were the principal training focus. Mandatory ARP4761 FHA processes require an assessment of pilot workload; both 737 MAX disasters occurred minutes after takeoff during initial climb out and relatively low altitudes: a high-risk area for aviation safety and pilot awareness. ARP4761 requires consideration of flight phase and pilot workload; had this been properly performed during the 737 MAX and MCAS FHA, the MCAS system response should have lessened the automatic horizontal stabilizer deflection angle while simultaneously alerting the pilot. Remember, automatic descent initiation is obviously problematic when the aircraft is still climbing out after takeoff – the ground (water) is just 40 seconds away instead of 5-10 minutes. This failure violated the “Consider Flight Phase and Pilot Workload during FHA Commandment”.

10.  Keep Certification independent: Maintain FAA or 3rd Party (Independent Designated Engineering Representative) oversight and proactive involvement instead of increasing Organization Design Approval (ODA) fraught with incestuous schedule/profit driven motivations.   Yes, this requires more FAA work meaning more funding (the annual FAA funding game is the laughingstock of the aviation world, trust me). Listen to the senior FAA personnel who specifically complained about the lack of oversight else such violates the “Independent Certification Commandment”.

11.   Apply mandatory ARP4761 System Safety Assessment specifying 10—5 (or 10—7 if active DAL B) Mean-Time-Between Failure (MTBF) to AOA sensors. Affirm AOA manufacturer’s MTBF calculations via screening and retest at airframe integrator. The 737 MAX sensors should not have experienced two separate failures on new aircraft. Such poor sensor reliability violates the “Prove Device Failure Rate MTBF Assumptions via Screening and Testing Commandment.”

12.  Keep things simple: aircraft are designed for the Average pilot, not the Best (military-trained) pilot. Forcing a pilot to learn to override a system via control stick updates, motor cutoff and trim adjustments is stupid (the Lion Air routine on the day’s prior flight with a well-trained pilot hitching a ride in the jumpseat). Instead, simply display “MCAS Activated – Continue or Disable?” on Primary Flight Displays with quick access for pilots to disable. Otherwise we violate the “Keep It Simple for Average Pilots Commandment.

13.  When the plane almost crashes on Sunday (Lion Air) it’s not readily cleared for safe flight the next day Monday without fixing the problem. Period. Aviation safety is built upon Root Cause Analysis and if you could not find the root cause you certainly could not then analyze it. This violates the “Don’t Be Stupid. Period. Commandment.”

There: Thirteen Aviation Safety Commandments. Interestingly, these Commandments are Forgiving: Had the 737 MAX events violated only one, or only seven, or only 12 of these Commandments, it’s my opinion the two crashes could have been prevented. But when thirteen of thirteen Commandments are broken, it’s time to seek Forgiveness – but only after going back and learning Prevention. Because after all, requesting Permission really is better than requesting Forgiveness.

Author’s note: AFuzion normally sells its “Applying ARP4754A” whitepaper and “Applying ARP4761” whitepaper for $50 each. But in the interest of world aviation safety, these two papers are free for anyone to download from now through April 30, 2019. Simply go to the Whitepaper page on www.afuzion.com. Download for free here: https://afuzion.com/rp-4761a-introduction-avionics-safety/

Safe Skies, Vance Hilderman, CEO, AFuzion Inc.