(Published with permission. Please address your comments to the author’s email address at bottom of the page.)
Copyright March 2017. All Rights Reserved by Charles J. Odendhal III
Preventing Jet Engine Failures
60-Year-Old Solution to 75 years of Jet Engine Failures?
Jet engine failures in aircraft can result when pieces of the engine, usually compressor blades, less commonly turbine buckets and disks, are cracked by “Metal Fatigue” and rupture, while bearing failures usually developed more slowly. Bearing failures can also tear blades loose and they may bind up inside engine housings, either tearing loose or twisting the aircraft off course. Some uncontained parts will often shoot out at supersonic speeds and damage the aircraft, even damage other engines. The jet engines on most modern transport aircraft are usually located on fuel wasting wing mounted pylons or externally at the rear to prevent uncontained engine failures from causing damage to the aircraft. They are held in place by only two “fused bolts,” which are designed to break apart under the excessive vibration created by engine failure and allow a failing engine to fall away from the aircraft. Jet engines mounted internally in the tail were used by the DC-10, Boeing 727 and Tu-154, but uncontainment still often results in serious damage to the aircraft.
Many uncontained engine failures mounted externally can and do tear away more or less safely, but some create sufficient airframe damage to result in aircraft destruction. To date, there has been hundreds of contained and uncontained jet engine failures with thousands of deaths.
Limited History of Engine Failures
Starting in 1942, German jet powered aircraft destroyed themselves at an alarming rate due to (uncontained) turbine section failures. Poor quality metal in turbine section buckets (blades) allowed failures to bind up the turbine section, rotating the aircraft, or tearing the engine and airframe apart; destroying the aircraft. Engineers either armored their engine’s housings to help contained a failure or mounted their engines below the wings on fragile bolts so an engine could fall away when it failed.
In 1952, the first turbojet powered commercial transport, the British DH-106 Comet Jetliner, had engines mounted inside its wings for more aerodynamic efficiency. Engine failures were a serious problem, but no crashes resulted. However, Comet engineers surrounded their engines with armor plating to prevent airframe damage from uncontained engine failures, but this added costly weight.
June 20, 1965, Pan Am Flight 843, 707-321B, had a failure in engine #4, right outboard, which came loose, but then tore off 25 feet of the right wing. The pilot maintained control and landed safely. This failure convinced Larry Booda of Aviation Week to support my appearance before a Senate Subcommittee on Aviation Safety regarding my failure warning system; which did not end well.
July 19, 1989, an uncontained failure of an engine mounted internally was the cause of the dramatic United Airlines DC-10 crash at the Sioux City, Iowa, airport. The center engine exploded inside the tail section and debris severed all three hydraulic lines, leading to a total loss of flight controls. However, skillful use of the two-remaining wing mounted engines saved many lives.
December 29, 1991, China Air Flight 358, 747-200, experienced a failure in engine #3, right inboard which tore the engine loose, but it also swung over and tore off engine #4, left inboard.
October 4, 1992, El Al 1862, 747-200F, lost both engines #3 and #4 on takeoff from Amsterdam airport. The crippled aircraft crashed into apartments, killing both crew and many residents.
July 17, 1996, TWA Flight 800, a Boeing 747-100, exploded 11 miles south of Long Island, NY; killing all 230 occupants. Conspiracy theorists immediately claimed this was evidence of “Friendly Fire” coming from a Navy vessel or a deliberate terrorist attack. Some witnesses reported seeing a missile rise from the ocean and explode next to the aircraft. Other witnesses reported seeing a “Streak of Light” falling from the sky. The investigation was extensive, with official “Cover-up” claimed by many.
After most of the aircraft was recovered, the National Transportation Safety Board (NTSB) reported the Center Wing (fuel) Tank had exploded and the NTSB soon produced a movie showing the aircraft nose blowing off, then, the main section climbing and falling intact; all four engines in place. However, this was not correct as engine #3 was found two miles short of the main debris field.
Disassembly of engine #3 disclosed it had suffered an uncontained failure, that could have sent debris into the right side of the 747, where an almost empty fuel tank contained explosive vapors. Engines #1, #2 and #4 were found miles away, intact. They were not disassembled.
Crew members in an approaching 737 aircraft reported seeing a bright light on Flight 800’s 747 (uncontained failure) followed by an explosion (Center Wing Tank) and then two fireballs going down. They did not see any missiles in the clear air and daylight. They were a few thousand feet away, while the other witnesses were 11 to 15 miles distant. Clearly another uncontained engine disaster.
November 4, 2010, Qantus Flight 32, an Airbus Super Jumbo A380-852, with Rolls Royce Trent 900 engines, experienced an uncontained engine failure of engine #2, left inboard, which damaged the nacelle, landing gear, flaps and ripped over a fuel tank. The fire started near engine #1, left outboard, but burnt out harmlessly, although it could have been worse. After a safe landing, fuel ran down onto the brakes and threatened to catch fire. Engine #1 was still running, with no controls so it had to be flooded with water from a fire hose. No one was injured although all could have easily died.
Recent Jet Aircraft Engine Failures
August 27, 2016, a Southwest Airlines Boeing 737-700, with GE CFM-56 engines, flying between New Orleans and Orlando. It experienced an uncontained engine failure in the left engine, during which the forward section of the nacelle was ripped off by a fan blade breaking loose and exiting the engine. A 5-inch by 16-inch hole was also found in the fuselage, just above the left wing. (Luckily the blade did not rupture a fuel tank and start a fire.)
The National Transportation Safety Board (NTSB) reported the cause was “Metal Fatigue.” It reads: “The fracture surface of the missing blade’s root showed curving crack lines consistent with fatigue crack growth.” These “Arrest lines” are like tree rings showing years of growth or, in this case, cycles of high stress; such as takeoffs. This is consistent with hydrogen embrittlement cracking.
October 28, 2016, a more recent example is the American Airlines’ Boeing 767, with GE CF6 engines, that was taking off from O’Hare Airport, Chicago, IL. Here, a cracked turbine disk is reportedly blamed for the uncontained material failure of the right engine. The disk exploded and debris tore open a fuel tank in the right wing, resulting in a fire that melted the wing. One piece of the turbine disk was found over 3,300 feet away. While a Kevlar shield had been installed to try to prevent uncontained debris escaping from the forward fan section, the adjoining wing had little chance of escaping from being penetrated by debris from other engine sections.
Fortunately, the pilot was able to abort the aircraft take-off, bring the plane to a stop near the end of the runway, and everyone escaped relatively unharmed. Had the aircraft become airborne, all the occupants most likely would have died in the inevitable crash.
November 14, 2016, an article in the Chicago Tribune indicated that engine failures were not uncommon with GE CF6 engines; with over 4,000 similar engines in service in 2017.
January 4, 2017, a USAF B-52 bomber flying out of Minot AFB, ND, suffered a “catastrophic internal engine failure” in one of its Pratt & Whitney TF 33-P-3/103 engines. The engine then ejected relatively harmlessly from the twin-engine pod and fell into a river bed about 25 miles NE of the base. An Air Force spokesman claimed this was a “one-off failure.” However, while investigating blade failures at Tinker AFB, OK, during 1958-60, I was personally aware of several uncontained engine failure reports on B-52 aircraft. It is obvious there were many more failures before 1958; which is why I was hired by the US Air Force. However, it seems the military does not usually report the results of contained or uncontained engine failures to the public, except when they are obvious to everyone.
All makers of turbojet engines have had their engines fail without warning, with many uncontained failures resulting in severe airframe damage, some with total aircraft destruction and fatalities. While “Metal Fatigue” is often considered the cause, I have good reason to believe that most engine failures are the result of Hydrogen Embrittlement.
A simplistic explanation of Hydrogen Embrittlement (HE) is that all metals are composed of crystals which can individually deform without damage if the amount of change is within their modulus of elasticity, even if excess hydrogen atoms are present inside their molecules. However, many, but not all, metals will absorb hydrogen atoms inside their molecules. Then, when stress is applied beyond their modulus of elasticity, usually by cyclic bending, the hydrogen atoms inside these molecules can migrate to areas where the crystals can be distorted beyond their modulus of elasticity. There the hydrogen atoms can collect between the crystals as hydrogen molecules and separate the crystal grain boundaries; effectively reducing the strength of materials, which can result in stress rupture and structural failure.
Hydrogen Embrittlement became part of my life in the 1950s. I started with a team researching material failures in drilling equipment at Shell (Oil) Development in Emeryville, CA. However, with no stress applied, the materials used in the drilling of oil wells would often pass all routine tests. Then, when subjected to vibratory stress, HE would cause an otherwise sound part to fail. We tried to determine how to rid equipment of the hydrogen atoms, but it soon appeared that measuring the amount of hydrogen atoms absorbed within materials during manufacturer was impossible.
Detection of HE in compressor blade failures
In 1958, I was hired by the US Air Force to assist in an investigation of foreign object damage (FOD), which was claimed to be the cause of most turbojet engine failures. These failures and often fatal crashes of Air Force aircraft were largely assumed by mechanical engineers to be caused by metal fatigue, initiated by FOD. I was assigned to determine how much FOD damage could be allowed in the leading edges of compressor section blades, to reduce the frequency of uncontained engine failures; since there appeared to be no way to eliminate them.
It became obvious to me that most blade failures were directly related to the duration of compressor stall or vibratory stress. This is where incoming airflow pressures varied in the engine inlet during take-offs and climb. A compressor stall would cause the blades most affected by airflow variation to vibrate the blades. Bombers seem to be most affected by such stress. Commercial aircraft also spend similar periods enduring compressor stalls, especially during take-off.
I secured relatively undamaged examples of blades from the overhaul area which were found to develop the most cracks in service and fitted them with strain gages. Then I mounted them in a base of dense metal and vibrated them in a high pressure airflow. When vibration frequency became a harmonic, the strain of the vibration would greatly increase and many blades cracked during tests, with no FOD present. In contrast, some blades with FOD beyond existing limits would endure testing stresses for longer periods without failure. Clearly, FOD was not as critical as engineers believed.
Engineers also insisted that the highest strain was in the leading and trailing edges, with only moderate strain in the concave surface and little on the convex surface. The areas where FOD nicks were considered the most critical was where my investigation was to be concentrated. However, I found many blades with cracks on their convex surfaces, located around 1/3 up from the base and 1/3 in from the leading edge, with no evidence of FOD. I began to suspect HE as the cause and inflight detection of cracking as a cure.
GE engineers, responsible for the engines being tested, informed me that I was wasting time and money since they had already determined the most critical areas of stress based on theory. Besides, they claimed there was no way to test blades under service conditions, using strain gages, in an operational engine. However, I found a Chicago company that made a slip ring which could be attached to the nose of the engine and transmit the necessary strain gage information. It appeared the engine manufacturers were not being truthful, so I requested the stress testing of strain gaged instrumented blades in an operational engine, using an (unused) Tinker AFB test cell.
I also found a material which would indicate the location and intensity of stresses during airflow testing. It was called Photostress Plastic and changed color in accord with the amount of strain. (Invented by Dr. Felix Zandmann of Tatnall Measuring Systems, Phoenixville, PA). This verified to me that the most critical area was located at 1/3 of span on the CONVEX surface. This also indicated to me that GE engineers had not fully tested blades for vibratory stress and the HE I was trying to prevent.
Suggested Engine Failure Warning Device
I wrote a paper about my observations and detailed how a simple device might warn pilots of impending engine failures. I sent it to all concerned, including the Wright-Patterson AFB Research Laboratory, Dayton, OH, and GE Engineering, Evansdale, OH; manufacturer of the engines then being overhauled at Tinker; J47, J57 and J79.
My invention, submitted as a suggestion, was a device which could be fitted to existing inspection ports, through which inside physical inspections are made in almost all engines. I believed that the outward movement of blade or bucket tips created by an initial material failure or bearing wear would be detected in-flight by blade tip contact with the device located in the engine housing and warn a pilot that a material failure was in progress, allowing for continued operation and safe landing.
There was no known interest shown on the part of the Air Force or GE engineers.
However, a secretary to our lab boss called me and asked that I take the contents of the office waste paper basket to the department paper shredder. She was aware of my interest in anything related to blade failures and the hostility of our lab boss to my “meddling,” since he felt that I should be restricting my investigation to the simple testing of nick depth to establish safer FOD limits.
I was puzzled by her request, not being a janitor, until she indicated there was something of interest to me inside the basket of waste paper. There I found a letter, possibly a copy, from a Mr. R. I. Brown, Supervisor of Engineering at GE’s Evansdale, Ohio, facility. He clearly stated that “GE Engineering had never done vibratory stress level testing on the blades in question.” He also suggested I was wasting my time and that should be “corrected”.
I wrote a report for the Air Force in 1959 detailing my belief that FOD may not be the primary cause of uncontained engine failures. I explained that many engine failures may have been caused by incorrect manufacturing processes which allowed excessive hydrogen atoms to be retained within the metal of rotating components; which could be demonstrated to create cracks under vibratory stress, leading to material rupture and uncontained engine failures. I stated that the real cause of uncontained engine failures was most likely HE; an industry problem still with us today.
I maintained that the then vacuuming of runways for FOD was probably unnecessary. I urged that more attention be paid to preventing failures by detecting the elongation of blades and buckets; many of which could be shown to rub the inside of housings long before failure. I suggested a change in the manufacturing procedures might reduce or eliminate HE. I also encouraged a change in aircraft operation to reduce compressor stalls.
However, implying that the engine manufactures’ engineers were the cause of engine failures was not well received. A fellow worker warned me that I might soon be terminated for researching other issues than FOD nicks. So, I resigned to do my own research.
Development of the Rotor Blade Monitor
To do this, I formed an Oklahoma stock company, the Oden Research Corporation, with my savings and money from investors, to research, design and test my concept in an operational engine. I also went to Dunlapp & Levy, Patent Attorneys, and hired them to do a Patent Search. They reported that the device was not patentable since the design had been patented in the past for a similar use with other rotating mechanical equipment.
While working for the Air Force, I noticed that blades being scrapped for various reasons, including some with tips showing heavy rubbing contact inside engine housings due to bearing failures, were then sent to a commercial salvage yard in Purcell, OK. I went to the salvage company and told them I was trying prevent uncontained engine failures. To my delight, the owner offered me an operational J-47 engine for me to equip and test with my proposed device and allowed me to sort through the thousands of blades obtained from the Tinker AFB overhaul facility; all free of charge.
I spent many hours sorting through hundreds of blades. I coated them with a special red dye which would soak into any cracks, which were not easy to see. I then cleaned off the dye with a solvent and re-coated the suspect blades with a white powder. The dye in any cracks would come to the surface and the red marks of a crack would become clearly visible, sometimes two or more cracks appeared; most of which were unrelated to any FOD.
Over a time, I acquired a fair number of blades that clearly showed cracks, with no FOD; mainly on the convex surface. Most of the blades indicated evidence of cracking in stages, much like tree rings, representing multiple take-offs and related stresses; suggesting the possibility of hours of flight time before failure. I also found blades which had heavy rub marks on their tips, that had come from engines which had bearing failures, indicating my device would have worked as designed either way.
How long a rotating component would last after blade elongation was first detected in flight was the question, although the cracked blades found in the salvage yard suggested it might be many flying hours; allowing sufficient time for a safe landing.
With finances being tight due to my limited income, I was not able to move the engine to my rural property and install the defective blades for testing. However, I did design a reinforced structure to house the engine, which would contain any possible engine explosion; in the event I got funding. My plan was to equip an operational engine with my “Rotor Blade Monitor” and run it to see if it could warn of a blade elongation before any potential engine failure.
Senate Subcommittee Hearing on Air Safety
In 1965, Larry Booda, a reporter with Aviation Week and Space Technology Magazine, became interested in my project and arranged for me to seek funding for my method to detect impending engine failures in a formal meeting with the Senate Aviation Subcommittee on Aviation Safety, in Washington DC.
I traveled to the nation’s capital and I was in awe of the impressive room, with senators and staff, plus military brass from every service and several engine manufactures’ representatives in attendance. I then presented my proposed invention and illustrated its operation with examples of defective blades which clearly showed a slow progression in cracking prior to blade failure. This was counter by the GE engineers’ conclusion, again supported by USAF engineers from Wright Patterson Air Development Center, that engine failures happened too quickly to be detected in a useful time. I considered this “No time to avoid failure” claim as a bogus assumption, but it was repeated many times over the years by GE and USAF engineers.
After the meeting, I was informed that the subcommittee would consider my proposal and my request for funding. However, while leaving the meeting, a young GE representative told me, “Your presentation made sense, but we will see that nothing happens.” Then another man demanded to know where I got the blades. He did not identify himself. I said, “The blades came from a Purcell OK Salvage.” He said, “We’ll checked it out. If they are found to be Air Force property, you could go to jail.”
Upon my return home, I learned that my request for funding had been denied and I was very discouraged. Then, the owner of Purcell Salvage asked me to meet him at Westheimer Field, where I based my airplane. He met me at dusk and asked that we walk over to my airplane; tied down some distance away on the flight line. (I was Commander of the local Civil Air Patrol Squadron.)
He led me to the side of the plane away from buildings and told me to look towards the airfield runway so our conversation could not be observed. He told me, “You really pissed off someone important. It would be best you never come back to the yard.” He implied it might not be safe for me.
Solution remains in limbo?
Therefore, without funding or access to an engine or blades for a demonstration, the Oden Research Corp was dissolved and my attempt to develop an aircraft engine warning system ceased. I eventually found work as an engineer for a British Leyland car distributor, investigating Jaguar, Austin and MG car defects in the western USA; based in Denver, Colorado.
It still seems inconceivable to me that hundreds of people with far more experience and greater qualifications have kept silent over the years. Surely, many FAA and NTSB investigators and engine manufacturers’ engineers must have known of the dangers of HE and its role in engine failures AND the need for a warning of impending rotor blade and disk failures. After all, my warning about what was causing rotating component failures and how my device might work to prevent them was sent to the Wright-Patterson AFB Research Laboratory in 1959, with a detailed explanation published in 1965.
Yet we still have engine failures and there seems to be little mention of HE in any accident investigation reports. One must wonder how far back HE was known to metallurgists and engineers as a possible factor in accidents caused by “Metal Fatigue,” so commonly blamed for accidents by the NTSB. I believe from first hand evidence that HE was a factor in USAF accidents at least 60 years ago, perhaps earlier, but everyone, except me, seems to have remained silent.
There have been recent efforts to protect the public and pilots from engine failure, but wrapping Kevlar Matting around just the fan section does not appear to be too successful.
What about the loss of engines, loss of aircraft and, more important, the loss of life? The loss of one engine might pay for the retrofit of hundreds of rubbing contacts. The loss of one aircraft might pay for a retrofit to an entire fleet of aircraft. The loss of life cannot be measured.
Present Situation, 2017
While my device may not work as designed, that has yet to be proven. Meanwhile, engine failures will continue to cause aircraft crashes worldwide, with the potential for an increased loss of life in a single event. The Rolls Royce Trent 900 engine has suffered several uncontained engine failures; thankfully without fatalities to date. One can only wonder what will happened when the next uncontained engine failure creates a related disaster in another Airbus A380, carrying up to 525 passengers, as it may sooner or later; with no warning system available.
Charles J. Odendhal III, 819 35th Ave., Greeley, CO 80634 970-371-0932 firstname.lastname@example.org
Administrator, Natural Hydrogen Energy, Ltd., odendhal@NH2E.com, Webpage: NH2E.com
Back to AERONAUTICS