Aircraft Accidents and Lessons Unlearned LXX: TransAir Flight 810

On July 2, 2021, at 01:10 Hawaii-Aleutian standard time, TransAir (Rhoades Aviation) flight 810 (TA810), a Boeing 737-275C (-200 series, ‘C’ designating cargo), registration N810TA, suffered a loss of power in the right – number (#) 2 – Pratt and Whitney JT8D-9A engine during climb out, after departing from Daniel K. Inouye International Airport (PHNL) destined for Kahului International Airport (PHOG). Following a series of operational errors, the aircraft was ditched in Mamala Bay. N810TA was one of five B737-200 aircraft in TransAir’s fleet. The National Transportation Safety Board (NTSB) assigned accident number DCA21FA174; the final report was released December 19, 2022. All information was pulled from the NTSB’s archives for this accident.

The NTSB’s probable cause for this accident: “The flight crewmembers’ misidentification of the damaged engine (after leveling off the airplane and reducing thrust) and their use of only the damaged engine for thrust during the remainder of the flight, resulting in an unintentional descent and forced ditching in the Pacific Ocean. Contributing to the accident were the flight crew’s ineffective crew resource management, high workload, and stress.”

Question: Did the #2 engine’s failure cause the pilots to become confused or did the pilots’ loss of control cause the #2 engine to fail? If one takes the probable cause as a gauge for how this accident investigation was conducted, the report couldn’t have been more off target. The pilots’ actions were a result of the #2 engine’s failure; the TA810 report should’ve addressed the engine problems first. If NTSB investigators paid attention to their own reports’ facts, they’d have seen the pilots’ actions were contributory; the pilots’ actions were only one symptom among many symptoms.

Why, then, did NTSB investigators dismiss the engine issues to focus on the not-as-prevalent issue of pilot communication? Consider this: if the engine problem was systemic – had occurred in other TransAir aircraft – wouldn’t it have been wiser to determine why the engine failed to prevent another engine from failing? The NTSB never investigated the engines’ histories properly. The question is moot because the NTSB wasted valuable time and opportunity in shrugging off the engines.

The accident aircraft suffered an engine malfunction during climb out … period! Nowhere in the probable cause did NTSB inspectors call attention to the #2 engine problem as a major contributor nor did the NTSB pursue any investigations into #2 engine’s last overhaul beyond posting the authorized release certificates in the Aircraft Maintenance/Human Performance (AMHP) Group’s Attachment 8. The NTSB didn’t explore any pertinent engine maintenance information. Why not?

Reviewing the AMHP Group Chairmen report, the maintenance investigator-in-charge (IIC) was an ‘aerospace engineer’. It’s unclear what engineering the aerospace engineer ‘engineered’. Did he/she design hydraulics systems? Anti-icing systems? Airliner lavatories? The IIC co-chaired with a human performance investigator who also lacked any aircraft maintenance experience or knowledge. The joint report was inconsistent; no plausible or useful conclusions were made. Interviews with Federal Aviation Administration (FAA) and Rhoades Aviation Inc. (RAI) were in Attachment 7 in the Docket.

The AMHP report demonstrated neither AMHP chairman brought any maintenance subject matter expertise or knowledge to the investigation. The AMHP report attachment 8, Engine Shop Records, had copies of the latest overhaul records for the accident aircraft’s two engines, but nothing described what the team found; what conclusions were considered; or if the overhaul records provided any clues to the engine failure.

The interviews provided nothing to help industry understand what may have happened. The term ‘monitoring’ or ‘trend monitoring’ was located 27 times in the interview transcript. However, the intent of engine trend monitoring or how it was accomplished at RAI couldn’t be understood. The terms were buzz words; expressions that sounded real maintenance-y – but nothing more. For example, on page 24, “So we’ll have trend monitoring comes up in the reports, how many cycles per plane in hours for the month, and then there’s raw parts or engine changes, it’s all in the CASS [Continuing Analysis and Surveillance System] report.”

The Assistant Director of Maintenance (DOM) spoke of trend monitoring, but it made no sense. Why? Because trend analyses draw data from different inputs, such as oil consumption, maintenance releases, hot starts, engine shutdowns, all followed by tracking programs for engine health reports. The operator tracks the data, or it’s contracted to an independent company. CASS is a surveillance system to monitor the ten elements of the air operator’s continuous aircraft maintenance program (CAMP) required per Title 14 code of federal regulations Part 121 Subpart L or Part 135 Subpart J. CASS doesn’t trend analyze engines. CASS tracks a CAMP’s efficiency which, per the interviews, didn’t appear very efficient. Meanwhile, both chairmen missed an obvious symptom.

On Attachment 7’s page 458, a lead mechanic (LM) was asked by RAI’s Director of Operations (DOO), “… shouldn’t our engine monitoring program, shouldn’t it have caught these engine problems we’ve had. And I said, well if there’s a trend established, you should see a spike and some sort of fuel flow or EGT [engine gas temperature]. And that I guess answered his question. He was satisfied with that.” Wait … What? To unpack this LM response to the DOO’s question the counter-questions would be: Why did the DOO ask an LM a question about data he or she had no access to? If the DOO did ask the question, why didn’t the LM say he wasn’t in possession of that information? To analyze trend analysis is not a lead mechanic function. LMs aren’t trained to analyze trends.

For the AMHP chairmen to ask the LM such a question was very naïve. They obviously didn’t understand the subject matter. Neither chairman asked effective questions; they didn’t pursue important resources or significant investigatory directions for finding out why the engine failed. Why didn’t the AMHP group interview whoever conducted RAI’s trend analysis? Why did they waste time interviewing people who did not collect or analyze data?

Just as disappointing was the overhaul facility, APECS Engine Center, the company that last worked on the accident aircraft’s engines, was never interviewed. The AMHP chairmen posted the work orders for the engine work in Attachment 8, but the interviews never showed APECS was contacted.

Interviews with RAI inspectors about engine borescopes would’ve been productive. Instead, the AMHP chairmen questioned an LM about borescopes, something the Inspection department does – not LMs. According to the LM, borescopes are performed frequently. Then on page 117, the Assistant DOM says, “No, we’re not --- we don’t do borescopes.” He even said, “We’re not qualified for borescope.” A qualified investigator’s follow-up questions would have been: Do you do borescopes or not? Who does them? – (because the Assistant DOM didn’t know). Did you borescope the accident aircraft’s engines? What turbine or compressor stages did you look at?

By far, the best reply was on page 399, when an LM was asked, “Did you get to the borescope?” The LM’s reply was, “No, we didn’t, since I didn’t find any damage externally.” What? His reply would be funny if it wasn’t so scary. Obviously LMs were not the source for borescope questions since they knew less than the Assistant DOM, like: What borescopes are for. A seasoned maintenance NTSB investigator would’ve known the questions to ask and of whom. A seasoned investigator would’ve pursued the inspector(s) who performed borescopes. The AMHP chairmen were out of their element.

On page 5 of the AMHP Chairmen’s report, Table 1 – Engine and APU Information chart, records show the #1 engine, serial number 674548, was maintained by APECS and released on September 19, 2019. The #1 engine was installed on the accident aircraft on October 28, 2019; the chart shows the #1 engine had accumulated 1055.2 (flight?) hours since the last shop visit and had 2085 (flight?) cycles since the last shop visit. The AMHP chairmen show 2085 flight cycles (takeoff to landing) but only 1055 flight hours. This meant for every hour flown, the accident aircraft took off and landed two times.

On the same chart, the #2 engine, serial number 657277, was also maintained by APECS, this time it was released April 5, 2019. The #2 engine was installed on the accident aircraft on September 20, 2019; the chart shows the #2 engine had accumulated 1055.2 (flight?) hours since the last shop visit and had 2085 (flight?) cycles since the last shop visit. Again, the AMHP chairmen show 2085 flight cycles (takeoffs and landings) but only 1055 flight hours. Is that even possible?

NOTE: A flight hour (FH) is generally defined as the time between engine start for the intent to fly and engine shutdown; it may or may not include maintenance runs. A flight cycle (FC) is generally defined as the initiation of the takeoff roll to completing the landing rollout.

What is interesting about these charted numbers is the #1 and #2 engine show the exact same number of flight cycles and flight hours since being installed on the accident aircraft, despite the fact the #2 engine was installed 38 days before the #1 engine. This would be possible if the engines were both installed during the duration of a heavy maintenance visit, but per page 7, Table 4 – Maintenance Checks chart, the aircraft wasn’t in maintenance check 38 days during the time period the two engines were installed. The chance both engines, installed 38 days apart, would have identical FHs and FCs was mathematically impossible. During engine start only one engine is started at a time; usually the same engine is started first, such as the #2 followed by #1 engine. According to Table 1, both engines had identical times down to the 0.10 hour … after almost two years. The AMHP chairmen missed this obvious discrepancy.

Table 1 – Engine and APU Information chart, the FH per engine versus FC, showed if each engine had, since the last shop visit – or installation – 1055.2 FHs over 2085 FCs, that means the average flight (FHs /FCs) lasted one half-hour, from engine start, taxi out, takeoff, climb out, cruise, approach, landing, taxi to the gate, to shut down – for two years. That’s a tight schedule over two years, if it was possible, even jumping from Hawaiian island to Hawaiian island. The AMHP chairmen didn’t question those numbers.

Then there was the ‘C’ (heavy phase) check schedule. As per Table 3 – Check Interval, each of the ‘C’ checks, from 1C to 6C, was separated by 4,000 flight hours or 24 months (whichever comes first?). Table 4 – Maintenance Checks showed the accident aircraft went through 4C, 5C and 6C simultaneous checks on February 7, 2015. The next ‘C’ check, a 1C check should’ve taken place in February 2017, not April 25, 2021, which was 74 months/2735.1 flight hours later. Then 1C, 2C and 3C were performed together. Was TransAir allowed to bundle their ‘C’ checks in this manner? Did they violate their CAMP? Why didn’t either AMHP chairman question this?

Table 4 Maintenance Checks - Dates accomplished

Both engines were recovered from the ocean floor on October 16, 2021, 106 days after the accident. The saltwater environment had a dramatic effect on the engine components, such as compressor discs, blades, and shafts. On page 16 of the AMHP factual report, the AMHP chairmen deferred to the Powerplant chairman (PWC) to answer questions about the engines, in relation to the accident: “Further Details on the engines can be found in the Powerplants Group Chairman Factual Report.” Why? Why dismiss investigation responsibility? The PWC doesn’t look into maintenance issues. The PWC position has the engineering data responsibilities to work with the AMHP to determine where the #2 engine’s upkeep went ‘off-the-rails’. The PWC also should’ve participated in the interviews.

The PWC found no damage beyond saltwater corrosion and blades that experienced a rapid deceleration caused by the turning engines contacting the water and the engine shroud compressing around the various discs. The Powerplant Group attachment 2 only contained ambiguous N1 and N2 data, but not much else. Who produced this health report, TransAir or a contractor? The AMHP group should’ve pursued this data dealing with maintenance planning, not leaving that to an engineering powerplant investigation group.

Pratt and Whitney (P&W) submitted a Metallurgical Investigation Final Report for the Powerplant group, which stated of blade 6 of the turbine section, “Based on these observations, it seems likely that internal oxidation/corrosion resulted in loss of load bearing cross section, which resulted in a stress rupture fracture.” P&W determined a good candidate for root cause of the engine failure, thus leading to the accident. The AMHP group should’ve been working with P&W before and after this finding, then investigating if the oxidation and/or corrosion was a result of poor maintenance; the accident; poor non-destructive inspecting; failure during the overhaul process; environmental or a combination of one or more of these and other contributors. TransAir’s remaining four B737-200 aircraft could’ve been subject to the same failures. Since the engine failure wasn’t given its due credit or investigatory attention, the P&W findings were ignored.

This NTSB accident investigation, specifically the AMHP investigation, failed to focus on priorities; this would be considered an amateurish mistake if the future ripple effects weren’t so serious and if it wasn’t supposed to be accomplished professionally. It is, however, another reminder the NTSB doesn’t consider aircraft maintenance seriously when investigating any aviation accident – not just major accidents – and the NTSB fails to accept the fact the FAA has an aviation safety inspector workforce that oversees the aviation industry, that half that FAA workforce is Airworthiness – aircraft maintenance and avionics.

The AMHP chairmen missed major findings. They treated terms like ‘borescope’ as if they sounded technically impressive to say and not as if they actually required understanding. The AMHP chairmen treated the term ‘trend monitoring’ with the same dismissive approach. Their interview questions were as if mentioning the words ‘borescope’ and ‘trend monitoring’ gave their questions an aircraft maintenance authenticity. However, they sounded … unqualified, because they didn’t know who to ask questions of. The AMHP chairmen’s failure to follow the FH, FC and ‘C’ check hours just added to the impression the NTSB didn’t take this accident seriously.

The fact the NTSB, after 50+ years, still assigns engineers and human performance investigators – those who have absolutely no understanding or experience in aircraft maintenance matters – to investigate aircraft maintenance issues will continue to result in accident investigations that miss major problems that lead to future accidents. Make no mistake, TransAir 810 may have been a small operator’s accident, but the NTSB’s complacent approach will rebound in other industry safety failures. This will become more obvious as the industry continues to squint as it comes out of the darkness of CØVID lockdowns. Metaphors and analogies aside, if this is an example of safety proactiveness, we’re in big trouble.