Air Safety Through Investigation JULY-SEPTEMBER 2019 Journal of the International Society of Air Safety Investigators

Developing a Photogrammetric Model for Wind Simulation page 4

Investigating a Unique UAV Gear Collapse page 12

Application of Innovative Investigation Technologies page 15 5 Millimeter Crack Leading to an Engine Fire page 18

ISASI Kapustin Scholarship Essay—Skydiving Operations and Safety Investigation page 26 CONTENTS Air Safety Through Investigation Journal of the International Society of Air Safety Investigators FEATURES Volume 52, Number 3 Publisher Frank Del Gandio Richard B. Stone 4 Developing a Photogrammetric Model for Wind Simulation Editorial Advisor Editor J. Gary DiNunno By Michael Bauer, William English, and Michael Richards, U.S. National Transportation Design Editor Jesica Ferry Safety Board, and Matthew Grzych, the Boeing Company—The authors noted that the Susan Fager activities conducted serve as a proof of concept for an innovative investigative technique, Associate Editor demonstrating the breadth of potential use cases for sUAS to support accident investigation. ISASI Forum (ISSN 1088-8128) is published quar- terly by the International Society of Air Safety Investigators. Opinions expressed by authors do 12 Investigating a Unique UAV Gear Collapse not necessarily represent official ISASI position By James Buse, Senior Technical Lead System Safety, and Jeff Kraus, UAS Flight Test or policy. Safety Lead, the Boeing Company—The authors describe their investigation of a landing Editorial Offices: Park Center, 107 East Holly Ave- incident during the first flight of a high-tech drone in June 2012 and the measures Boeing nue, Suite 11, Sterling, VA 20164-5405. Telephone took to correct the problem. 703-430-9668. Fax 703-430-4970. E-mail address, [email protected]; for editor, jgdassociates@ starpower.net. Internet website: www.isasi.org. 15 Application of Innovative Investigation Techniques ISASI Forum is not responsible for unsolicited By Sundeep Gupta, Accident Investigator, Product Safety, Airbus—TThe author describes manuscripts, photographs, or other materials. innovative investigation technology, including satellite imagery, ballistic analysis, radar Unsolicited materials will be returned only if scanning, and 3-D laser scanning used to locate engine components scattered from an submitted with a self-addressed, stamped enve- lope. ISASI Forum reserves the right to reject, uncontained engine failure on an A380 cruising at 37,000 feet over Greenland. delete, summarize, or edit for space con- siderations any submitted article. To facilitate editorial production processes, American Eng- 18 5 Millimeter Crack Leading to an Engine Fire lish spelling of words is used. By David Lim, Principal Investigator, Transport Safety Investigation Bureau, Singapore— The author presents lessons learned from investigation of an engine fire on a B-777-300ER that Copyright © 2019—International Society of Air departed from Singapore to Milan in June 2016. Safety Investigators, all rights reserved. Publica- tion in any form is prohibited without permis- sion. ISASI Forum registered U.S. Patent and 26 ISASI Kapustin Scholarship Essay—How an Extreme Sport T.M. Office. Opinions expressed by authors do not necessarily represent official ISASI position Highlights Broader Issues for Air Safety Investigators: Sky Diving or policy. Permission to reprint is available upon Operations and Safety Investigations application to the editorial offices. By Avery Katz, Embry-Riddle Aeronautic University, 2018 ISASI Rudolf Kapustin Memorial Publisher’s Editorial Profile: ISASI Forum is print- Scholarship Recipient—The author discusses how an extreme sport presents unusual and ed in the United States and published for profes- broader issues that accident investigators must address. sional air safety investigators who are members of the International Society of Air Safety Inves- tigators. Editorial content emphasizes accident investigation findings, investigative techniques and experiences, regulatory issues, industry ac- DEPARTMENTS cident prevention developments, and ISASI and member involvement and information. 2 Contents 3 President’s View Subscriptions: A subscription to members is pro- 28 News Roundup vided as a portion of dues. Rate for nonmem- 30 ISASI Information bers (domestic and Canada) is US$28; Rate for 32 ANZSASI Holds Annual Regional Seminar nonmember international is US$30. Rate for all libraries and schools is US$24. For subscription information, call 703-430-9668. Additional or replacement ISASI Forum issues: Domestic and Canada US$4; international member US$4; do- ABOUT THE COVER mestic and Canada nonmember US$6; interna- tional nonmember US$8. A Boeing Phantom Eye UAV liquid-hydrogen-fueled prototype lifts off from a lakebed at Edwards Air Force Base in California, USA, in June 2012 for its first flight. The high-altitude, long-endurance aircraft is designed to remain airborne for up to four days while reaching altitudes of up to 65,000 feet. Upon landing, the nosewheel separated as it contacted the runway, leading to a unique inves- tigation that required unusual procedures (see page 12). INCORPORATED AUGUST 31, 1964 2 • July-September 2019 ISASI Forum PRESIDENT’S VIEW

CONTINUING TO IMPROVE FLIGHT SAFETY IN THE YEARS TO COME

WE’RE PROUD OF OUR HISTORY, OUR PRESENT EFFORTS, AND OUR WILLINGNESS TO PROJECT OUR EXPERIENCE AND EXPERTISE INTO AIR SAFETY’S FUTURE.

s I write my opening to offer aviation students a of interesting air safety distributed to ISASI 2019 statement for ISASI scholarship that allows them professionals who are just participants and then will 2019 in early Sep- to attend our seminar and starting their careers. This be posted on our website. A tember, I think about other training magazine format is sent It’s safe to say that ISASI, many of the programs our programs. We’re a sponsor of electronically to members through our members, has Society has to offer to ISASI the Harry Robinson Fellow- and subscribers who prefer some presence everywhere members and to air safety ship program that provides to receive their information in the world where air safety throughout the world and the a stipend to an advanced through computers or other is an issue. We’re proud of efforts ISASI makes in prepa- college degree student who’s electronic devices, rather our history, our present ration for the future. Yes, our conducting research in an than a printed and mailed efforts, and our willingness premier event is the annual air safety subject. copy. Receiving either to project our experience seminar that gives air safety We’re an active participant format has two require- and expertise into air professionals an opportu- as an official observer organ- ments: You must be an ISASI safety’s future. That future nity to share their expertise ization in International Civil member in good standing may take us farther than through tutorials, technical Aviation Authority (ICAO) and your mailing or e-mail this world. The future of air presentations, and informal proceedings and rulemaking. address must be accurate. safety investigation depends networking discussions. But ICAO is the global aviation If you need to “fix” either of upon those of us who have there is a great deal more oc- authority within the United these requirements or wish experience to pass the torch curring throughout the year. Nations. For the last five to switch your magazine de- to next-generation air safety We recently participat- years, our Society represent- livery option, please contact professionals who can ed in the 54th Reachout atives have been a part of ISASI’s editorial offices (see assimilate what we learned program that provided air the ICAO Accident Investi- page 2 of this issue). and practiced while adding safety instruction not only gation Panel that meets an- During the spring new expertise and experi- to ISASI members in Paki- nually in Montreal, Quebec, ISASI International Council ence to continue improving stan, but also to government Canada. The work of the meeting, a committee was air and space flight safety in officials and personnel, group leads to updates and established to review the the years to come. members of the military, additions to ICAO standards Society’s official Positions and other safety profession- and procedures for accident on Air Safety Issues, which is als. There have been 3,272 investigations throughout posted on our website under participants in Reachouts the world. the Guidelines tab. This since this program’s incep- Our recent inauguration periodic review is designed tion in 2001. of a digital ISASI Forum is to ensure that our policies We encourage aviation an attempt to better reach and positions are current. students to learn about air members and subscribers The committee determined safety careers through our who are often away from that no changes are required student outreach programs their homes or worksites at this time. that include student ISASI and may want to read our A pamphlet about the chapters at accredited avi- magazine in a more timely Society’s history and Frank Del Gandio ation colleges and continue manner and also as a means accomplishments will be ISASI President July-September 2019 ISASI Forum • 3 DEVELOPING A PHOTOGRAMMETRIC MODEL FOR WIND SIMULATION By Michael Bauer, William English, and Michael Richards, U.S. National Transportation Safety Board, and Matthew Grzych, the Boeing Company

uring the on-scene investigation of an Incomplete meteorological data accident in which a transport-catego- for airport ry airplane overran a runway, NTSB The accident occurred on March 8, 2017, Dinvestigators found that the inboard at Willow Run Airport (YIP) in Ypsilanti, linkage components of the right elevator’s Michigan, USA, and an NTSB meteorologist geared tab were locked in an overcentered was assigned to the investigation. Official position and bent outboard, effectively wind observations at YIP were taken from an jamming the right elevator. This condition automated surface observing system (ASOS) prevented the right elevator from respond- anemometer located near midfield, 1,490 ing to cockpit control inputs and precluded meters (4,888 feet) east of the location where airplane pitch response and rotation during the airplane was parked (see Figure 1). This takeoff. Initial stages of the investigation anemometer was at a height of 10 meters Michael Bauer found no evidence that the condition result- (33 feet) above ground level (AGL) and was ed from improper maintenance or collision sited to observe unobstructed wind flow with a vehicle or other object. Before the from any direction. Although the ASOS is an accident flight, the airplane had been parked authoritative source of wind data, due to a during a period of abnormally high, gusting regional power outage the ASOS anemometer wind conditions, and flight data recorder became inoperative about four hours before, information (which recorded when the and remained inoperative until well after the right elevator was last in a different posi- time of the accident. Prior to the power loss, tion) showed that the right elevator became the ASOS anemometer’s maximum observed jammed sometime during the two days that wind gust (five-second average) was 55 knots the airplane was parked on the ramp before from the west-southwest. ASOS anemometer the accident. data did not distinguish between the hori- zontal and vertical components of the wind. Although two additional anemometers William English Statement of the problem were present at YIP, neither provided A lack of local wind data availability and the investigators with authoritative wind data likely effects of a large hangar on the wind for the accident day. A stand-alone weather characteristics where the airplane had been sensor (SAWS) that included an anemometer parked challenged the investigation’s ability was on the field northeast of the airplane’s to determine how the right elevator’s geared parked location; however, the SAWS became tab linkage became locked overcenter. Al- inoperative about the same time as the ASOS though a CFD wind simulation could provide anemometer, and archived data from before insight into the likely characteristics of the the power outage was not available. wind flow at the airplane’s parked location, Although another anemometer that was an accurate 3-D model of the large hangar owned by the local airport authority and upwind of the parking ramp was needed but located southeast of the airplane’s parked could not be feasibly produced using conven- location continued to provide data even after Michael Richards tional techniques. the ASOS and SAWS lost power, a compari- The wind flow information was critical for son of this anemometer’s data with the ASOS supporting the investigation’s development data (for times when the ASOS was opera- of a series of static and dynamic elevator tional) revealed that the airport authority’s load tests to determine what conditions, anemometer data showed a significant bias. consistent with the known circumstances Further, the airport authority’s anemometer of the accident, could result in the jammed provided data only every five minutes with elevator condition. Although the NTSB’s UAS some unknown sampling criteria, was at a Team had been using a fleet of sUAS as a new height of about 2.7 to 3.1 meters (nine to 10 investigative technique for capturing image- feet), may not have been well sited, and had ry at various accident sites for approximately unknown maintenance standards. The issues a year, supporting the development of a 3-D associated with these two anemometers, model for use in a CFD wind simulation was combined with the power loss to the ASOS, Matthew Grzych a new application of this technology. left investigators without credible wind

4 • July-September 2019 ISASI Forum (Adapted with permission from the authors’ technical paper ti- DEVELOPING A PHOTOGRAMMETRIC MODEL FOR WIND SIMULATION tled Use of sUAS in Developing a Photogrammetric Model for Wind Simulation presented during ISASI 2018, Oct. 30–Nov. 1, By Michael Bauer, William English, and Michael Richards, U.S. National Transportation Safety Board, and Matthew Grzych, the Boeing Company 2018, in Dubai, the United Arab Emirates. The authors noted that this white paper is not intended to be a scientific study. The activities conducted served as a proof of concept for an innovative investigative technique, demonstrating the breadth information for YIP for about a four-hour the likelihood of a digital of potential use cases for small unmanned aircraft systems to period before the accident. 3-D model of it already support accident investigation. The theme for ISASI 2018 was existing—especially with “The Future of Aircraft Accident Investigation.” The full pres- Potential effect of large hangar on the fidelity necessary entation can be found on the ISASI website at www.isasi.org in for the CFD work—was the Library tab under Technical Presentations.—Editor) wind flow remote. Creating a 3-D Although the ASOS anemometer provid- ed reliable, authoritative airport wind information while it was operative, the airplane had been parked facing north immediately downwind, on the east side, from a large hangar. The hangar was more than 0.4 kilometers (0.25 miles) long and equipped with duct work, chimneys, lad- ders, and detailed architecture; it was also surrounded by bushes, trees, and varied terrain near a retention pond. All of these characteristics would likely disrupt wind flow. Due to the size and characteristics of the hangar obstruction and surround- ing terrain, as well as the distance of the airplane’s parked location from the ASOS, investigators suspected that any available airport wind data from the ASOS likely did not represent the localized airflow at the airplane’s parked location (see Figure Figure 1. Airport map showing the relative locations of the hangar (dimensions highlighted 2, page 6). in blue), the accident airplane’s parking spot, and various weather observing equipment.

Need for CFD wind simulation model of the complex building and ter- tems investigator assigned to the investi- Given the lack of complete wind data for rain environment would conventionally gation approached the NTSB UAS Team YIP in the hours before the accident and require using blueprints, photographs, program lead to determine the feasibility the size and characteristics of the large terrain data, and CAD software to manu- of using the NTSB’s sUAS fleet to capture hangar upwind of the airplane’s parked lo- ally create the model. This would involve the necessary imagery and use it to create cation, determining the localized airflow considerable resources in time and a photogrammetric 3-D model of the that likely affected the airplane required manpower. For example, it would be dif- hangar. At the time, the NTSB UAS Team an additional investigative technique. ficult to manually create the basic hangar had been using drones for about one year The Boeing Company, which was a party structure alone, and adding the intricate to capture imagery at various accident to the NTSB’s investigation, was capable accessories (duct work, chimneys, lad- sites and had been using the commer- of performing a CFD wind simulation of ders, detailed architecture, bushes, trees, cially available photogrammetry software conditions downwind of hangar using terrain near the retention pond, etc.) to develop 3-D models of the sites from available wind data, pavement tempera- could take weeks of additional labor. Fur- the drone-captured imagery. Past models ture information from an airport author- ther, such a monumental effort still might developed using the software were well ity sensor, and additional meteorological not provide the level of accuracy needed. within the accuracy requirements spec- data. However, to maximize the fidelity of For effective use in the CFD simulation, ified to support the CFD simulation for the wind simulation, a photogrammetric which has 0.5 meters (1.6 feet) resolution, this investigation. 3-D model of the hangar that could be the input model must have at least that imported into the CFD simulation was level of accuracy to model small-scale Evaluating feasibility of using sUAS needed. turbulent wind patterns caused by the obstructions. imagery to develop 3-D model Determining photogrammetry product Limitations of traditional investigative compatibility techniques for developing 3-D hangar Basis for considering use of sUAS as Before accepting the mission, the NTSB model new investigative technique UAS Team needed to determine if the Due to the 1940s vintage of the hangar, The NTSB meteorologist and NTSB sys- output products from a photogrammetry

July-September 2019 ISASI Forum • 5 optimistic that such 3-D mesh objects could contribute greatly to the accuracy of the model environment.

Assessing size of structure and area to be modeled Once the initial feasibility of providing a CFD-compatible model was demonstrat- ed, the Boeing CFD specialist provided the UAS Team with specifications for the area to be modeled to support the wind simulation study. Conceptually, modeling the hangar was feasible for the UAS Team; however, at more than 0.4 kilometers (0.25 miles) long, the structure was much larger than anything the team had mod- eled before. In addition, the CFD special- ist requested mapping of a large area of the surrounding terrain. A map of the area of interest to be modeled was provided, with the highest priority being the southern portion of Figure 2. View looking west toward the hangar, taken from the approximate parked loca- the hangar, parking ramp, and a nearby tion of the accident airplane. (Airplane in picture is much closer to the hangar.) retention pond, with some lower pri- ority “reach” areas across a vehicle lot. program could be imported into the CFD developer of the commercially available Although the CFD specialist initially program used by Boeing. Photogram- photogrammetry software. The drone was requested a model of about only half of metry, or more specifically stereo photo- flown in a double grid pattern with the the hangar, the historical aspects of the grammetry, determines the 3-D coordi- camera pointed in combinations of nadir hangar, which was scheduled for demoli- nates of points on an object by employing (straight down) and slightly off vertical, tion, were considered. Known as Hangar measurements made from multiple as well as oblique patterns around the 1, it was one of the last remaining build- photographic images taken from different perimeter of the shed. The flight resulted ings of the former Willow Run Bomber positions. Common points are identified in about 70 photographs, which were Plant complex that produced the Consol- on each image, and the intersection of processed into a point cloud, and a 3-D idated B-24 Liberator bomber from World the lines from the camera locations to the mesh in .obj format was exported for use War II. The UAS Team determined that point on the object determines the 3-D lo- in the CFD simulation model. (A .obj file acquiring the imagery to model the entire cation. This 3-D point cloud model of the is a geometry definition file. It is an open hangar would not require substantially subject is then used to build a textured format and was developed by Wavefront more flight time. Thus, the UAS Team -de mesh representation of the object. Al- Technologies for its animation software. termined that mapping the entire hangar, though not as precise as the point cloud, The format is also used by various other thereby preserving it in a digital format, the mesh object could potentially be 3-D graphics applications.) would be worth the additional effort. used as a solid object in various software The Boeing CFD specialist was able to Further, the additional imagery would be environments, such as the CFD wind sim- import the test item into the CFD simula- available to the investigation, if required ulation, and provide more detail than the tion model and, based on the results, was for the analysis. simple geometric shapes normally used. To capture imagery and create a mesh representation of an object that could be tested with the CFD software, the NTSB UAS Team conducted proving flights at the test facility near the NTSB Train- ing Center in Ashburn, Virginia, USA. (This “test track” was the former George Washington University solar vehicle field, which the NTSB can use under agree- ment for UAS training and research and development.) During one proving flight, the UAS Team thoroughly captured images of an equipment shed at the field using flight patterns recommended by Pix4D, the Figure 3. Orthomosaic image showing planned sUAS mapping area in blue.

6 • July-September 2019 ISASI Forum Conducting sUAS flight missions Preflight planning to capture optimal imagery Considering the scope of the planned project, the NTSB UAS Team worked with specialists from Pix4D to ensure that the planned missions would be flown a manner that most effectively and effi- ciently gathered the necessary data. The UAS Team developed a flight plan that incorporated double grids over the top of the entire hangar, followed by two orbits at different altitudes and camera angles to ensure full coverage of the sides of the building. The altitudes of the grids and oblique flight paths were planned to give the best balance of coverage (including overlap) and accuracy. Grids flown at a higher alti- Figure 4. Sectional showing YIP and the surrounding airspace. tude provided better overlap and match- ing between the photographs for the officers were among the first responders eration, the eCOA was developed and photogrammetry processing, especially to the accident and used their sUAS to issued with ample lead time before the considering the somewhat homogenous photograph the wreckage, ground scars, mission. The NTSB UAS program lead nature of the hangar roof. In a general and runway.) Meanwhile, the NTSB UAS provided a detailed map of the operating sense, however, some photographic detail program lead began coordinating with area and proposed flight plans, and FAA is lost with higher altitudes because the the U.S. Federal Aviation Administration headquarters personnel coordinated with ground sample distance (GSD) increases (FAA) for airspace access. YIP air traffic control (ATC) personnel to as the camera moves further from the During this process, the UAS Team develop procedures for the flights. These ground. (In a digital photograph, GSD learned that the MSP was unavailable at combined efforts determined that opera- is the ground distance represented by the proposed times for the flight; thus, tions under the eCOA were subject to the the distance between the centers of two an eCOA was needed. (Note: Currently, following requirements: pixels.) the FAA provides an automated airspace • A flight team consisting of a remote The UAS Team flew the hangar grids at access system, low altitude authorization pilot-in-command (RPIC) and visual 46 meters (150 feet) AGL and positioned and notification capability, as well as observer (VO), the special government interest proto- the oblique flights to fill the frame to the • Issuance of a notice to airmen extent possible with the target object col, both of which allow for much more (NOTAM) prior to the day’s sUAS (hangar) and minimize the background, expeditious sUAS access to controlled flights (as coordinated between the such as sky and ramp. The team posi- airspace.) FAA and NTSB UAS program lead), tioned additional grids over the terrain The NTSB UAS Team had previously • Restrictions that sUAS operations used eCOAs for conducting sUAS oper- and a small hangar to the south and east must remain clear of all runways and ations over active accident scenes, but of the main hangar (see Figure 3). Pre- moving aircraft, and planning the flight missions resulted in an none of those site areas had been as large estimate of more than one hour of expected as the hangar mapping project, which • RPIC communication with the ATC flying and more than 1,000 images. required lengthy flights in close proximity tower personnel via two-way radio to manned aircraft operations. YIP was prior to each sortie. Airspace access considerations a busy freight hub in Class D airspace, The eCOA covered the planned mission At the time of the flights, gaining ap- underlying a shelf of the Detroit Class B date and included a contingency day to proval for conducting sUAS operations airspace (see Figure 4). account for weather or data quality/oper- in controlled airspace (like YIP) involved Further, the hangar was located in be- ational issues, if needed. either a time-consuming manual author- tween the ends of the two main runways ization process under Title 14 Code of and contained the U.S. Customs facility, Federal Regulations (CFR) Part 107 or a fixed-base operator (FBO), and other Mission considerations: weather, the use of a government agency public businesses. An active general aviation airport activity, and other challenges safety emergency Certificate of Waiver flight school was immediately south of, Weather or Authorization (eCOA). Initially, the and sharing ramp space with, the hangar. The mission day dawned with high, thin UAS Team coordinated with the Mich- The sUAS flights would need to operate overcast and light surface winds, which igan State Police (MSP) to ascertain if concurrently with manned aircraft flights were ideal for photogrammetric flights. the MSP’s jurisdictional Certificate of as the airport continued normal opera- However, the forecast called for storms to Waiver or Authorization (COA), which tions throughout the mission. move through by the middle of the day, covered the airport, could be used. (MSP Fortunately, with excellent FAA coop- leaving a relatively narrow time window July-September 2019 ISASI Forum • 7 to accomplish the flights. In considera- and then proceeded with the mission. communicated with each other, and tion of the approaching weather, the UAS This was the only (and very brief) tech- the RPIC responded by maneuvering Team prioritized the sorties to ensure nical issue of the day, which presumably the drone away from the grid. Although that the most critical portions of the occurred when a radio equipment user the Cessna and the drone never entered hangar were captured first in the event transmitted just as the drone was close to hazardous proximity to each other, the that subsequent sorties would have to be the tower. traffic scenario highlighted the value of canceled. close coordination and communication Airport ramp and traffic activity between the RPIC and VO. Preflight coordination Normal airport operations continued UAS Team members later asked ATC The UAS Team, consisting of an RPIC throughout the morning, which included personnel if they could see the drone and VO, arrived early to the hangar multiple cargo and business jet flights when it was flying. The tower controller and contacted the ATC personnel, who and a nearly constant flow of training stated that they had been looking for it on acknowledged the NOTAM information flights. During the sUAS sorties, airport each sortie but never saw it and also not- and added an advisory on the automatic ramp activity east of the hangar and near ed that no pilots reported seeing it. The terminal information service (ATIS) to the flight school hangar included air- Cessna pilot also stated that he did not advise local pilots of the UAS activity near planes entering the ramp for parking and see the drone. This information regarding the hangar below 61 meters (200 feet) passengers exiting and entering the build- the difficulty (or inability) of ATC person- AGL. Airport management was support- ings. To avoid overflight of vehicles and nel and pilots of manned aircraft to see ive, providing access to the airside ramp persons, the RPIC performed occasional drones served as a good reminder to the east of the hangar and the area around brief stand-downs by hovering to a safe RPIC and VO regarding the importance the flight school hangar. area or landing until the activity cleared. of maintaining required VLOS with the During one sortie that launched drone and communication with each oth- Mission sorties from the southwest corner of the hang- er, ATC personnel, and, if needed, pilots The UAS Team staged the drone for the ar toward a planned grid to cover the of manned aircraft when operating in an first sortie at the northwest corner of the retention pond abeam the approach end active airport environment (see Figure 5). hangar. Per the established plan, the first of Runway 5L, a pilot of a single-engine By the time that the last planned sortie overflight grid was flown at 46 meters Cessna executed a practice engine-out was completed, the storms were moving (150 feet) AGL. Breaking up the grid into landing and turned directly toward the very close to the airport. The desired are- four separate flights (one for each quad- as were covered, so after one last quality runway numbers, cutting the corner off rant) aided in maintaining required visual check the team packed up and headed in- the typical traffic pattern. The Cessna’s line of sight (VLOS) with the drone and side the FBO just in time to avoid the rain. tighter pattern could have brought the provided the opportunity for the team airplane concerningly close to the sUAS to quality check the image files while the flight operations area; however, both the Equipment used drone was landed between flights. RPIC and VO saw the Cessna turning, To maximize efficiency, the RPIC and sUAS, cameras, and DGPS VO worked a rotating plan of action: While the RPIC (UAS program lead) input the flight plan grids or orbits on the sUAS control tablet, the VO used a differential global positioning system (DGPS) to sur- vey ground control points (GCPs) around the hangar. After each sortie, while the RPIC swapped out drone batteries and prepared for the next mission, the VO downloaded the images from drone’s data card, reviewed them, and created backup copies of the images. As the missions proceeded, the team decided to fly an extra orbit around a small tower to ensure the best detail coverage of the complex structure and angles, which might affect the wind sim- ulation. While orbiting the small tower, which housed radio equipment for the U.S. Customs and FBO operations, the RPIC experienced a brief global posi- tioning system (GPS) and compass error with the sUAS. To address the issue, the RPIC immediately backed the drone away Figure 5. Image captured by the drone during the mission. (Note business jet on approach from the tower, which cleared the error, to Runway 5.)

8 • July-September 2019 ISASI Forum All equipment used for the effort was centimeters (1.6 to 3.9 inches). (The data commercially available, off the shelf. The sheet for the Trimble Geo 7 series units team brought two drones: the primary describes the unit’s accuracy and relia- drone used for the mission and one to bility and the factors that may introduce have available as a backup. The primary anomalies.) Thus, during the first step in aircraft was a DJI Phantom 4 Pro (see the processing, NTSB UAS Team mem- Figure 6) equipped with an FC6310 bers manually fine-tuned the geographic camera using the Sony Exmor one-inch location information by viewing the imag- (2.5-centimeter) complementary met- es containing the GCPs taken with the al-oxide semiconductor (CMOS) sensor, DGPS and marking the pixel representing which provided still photo resolution of the center of each GCP. 20 megapixels in .jpg or .raw format. The Following the marking of GCPs, the camera was equipped with a mechanical 3-D point cloud, consisting of millions shutter, which was ideal for use in cap- of data points, was constructed. Relative turing images via drone to support a pho- accuracy (in other words, measurements togrammetry application. (Mechanical within the point cloud) was measured to shutters avoid introducing image errors approximately 2.5 centimeters (1 inche). and distortions that can result from the (This accuracy is a function of the camera use of an electronic rolling shutter from a distance and focal length, amount of moving drone.) image overlap, and other factors.) A trian- The backup aircraft was a DJI Inspire gulated 3-D mesh was then built from the 1 equipped with the X3/FC350 camera point cloud. The mesh processing, which interpolates between points, required using the Sony Exmor 1/2.3-inch CMOS Figure 6. The RPIC flying a Phantom 4 Pro sensor, which provided a resolution of some iterative interaction to remove drone during a mission sortie. 12 megapixels. A Trimble Geo 7X DGPS artifacts in the mesh. Typical artifacts in was used to record the GCPs, which were the mesh were evident; for example, the of Democracy.” corrected using the continuously oper- color and texture of the hangar roof were Looking at the historic photos, it was ating reference station (CORS), which visually similar to the color and texture clear to the team that the exterior of was coincidentally located less than 0.8 of the paved ramp areas nearby, which re- the building was essentially the same kilometers (0.5 miles) away, in front of the quired some manual cleanup of the cloud as it was in 1944 . Hangar 1 was used airport fire station. to build the best 3-D mesh model. during World War II as a maintenance Two separate final mesh files (one of depot, providing storage for completed Software and processing the main hangar building structure and B-24s and modified aircraft (such as To build the 3-D model, the data was the other of the surrounding terrain, radar-equipped “Pathfinders”) until ready processed using the Pix4D Mapper soft- trees, brush, vehicles, and other nearby for military deployment. ware with a laptop PC compatible with structures) were exported as a 3-D model According to the YIP airport master the suggested specifications from Pix4D. in .obj file format and provided to the plan, Hangar 1 was scheduled to close Creating a detailed photogrammetric 3-D CFD wind simulation specialist at Boeing. and eventually be demolished on Nov. model of an object the size of the hangar 30, 2018. The only historic portion of the and surrounding terrain at the required Willow Run Bomber Plant complex to resolution required about 1,000 images. Digital preservation of historically remain (although not in original config- The process of developing the 3-D significant hangar uration) was the adjacent portion that model through Pix4D began with im- Initially, the Boeing CFD specialist asked served as the end of the assembly line, porting the photographs and validating for imagery of only the southern half of where aircraft were fueled. (At the time of the image information. During the first the hangar. However, as the NTSB UAS this report, that building was under reno- step, the software looked for common key Team members prepared for the mission, vation to house the museum.) When the points throughout the photographs and they learned the historical significance of NTSB UAS Team learned that this piece of marked thousands of overlapping points. the building. Known as Hangar 1, it was history was endangered, the initial plan Although photographs taken using the one of the last standing buildings of the to model only about half of the hangar drone have geographical information Willow Run Bomber Plant complex that changed: Half wasn’t good enough, so we embedded in the metadata, the accuracy produced the Consolidated B-24 Liber- decided to model the whole thing! As a re- is limited to that of an uncorrected GPS— ator bomber during World War II, and it sult, this historically significant building typically within five to 10 meters laterally housed part of the Yankee Air Museum was preserved in a digital format through with height information that can be very collection. The Willow Run factory com- the imagery collected and photogramme- inaccurate. The use of GCPs substantially plex was the first plant to build bomb- try output produced by the team. Fur- increased accuracy. Past experience of the ers on an assembly line, and it was the ther, obtaining the additional imagery to UAS Team with the Geo 7X showed that, home of the original “Rosie the Riveters,” model the entire hangar did not require under good conditions with nearby CORS, who were part of the unprecedented substantially more flight time, and the positional accuracy (relative to Earth’s expansion of women in the workforce, complete model was available for use in datum) could be as good as four to 10 building part of the American “Arsenal investigative analysis.

July-September 2019 ISASI Forum • 9 plane with the turbulent gusts and waves that were being generated downstream of the hangar. Figure 7 and Figure 8 show visualizations of select wind simulation results.

Full-scale elevator load testing As described previously, the inboard linkage components of the right elevator’s geared tab were locked in an overcentered position and bent outboard, effectively jamming the right elevator and prevent- ing airplane pitch response and rotation during takeoff. When the airplane is Figure 7. The 3-D visualization of wind simulation results for a discrete time showing parked on the ground, the elevators are turbulence generated downwind of the hangar. Note: The accident airplane’s elevators can free to move independently within the be seen on lower right side of image. confines of the upper and lower me- chanical stops, which inhibit elevator CFD wind simulation in the mesh to represent the parked loca- movement in each direction. Although Further processing to optimize 3-D hangar tion of the accident airplane. the elevator is designed to distribute model increased loading via a torsion bar and Results of CFD model analysis study The Boeing CFD specialist found that stop arm assembly when the elevator is in The Boeing CFD specialist used the final before the 3-D model mesh files pro- contact with a stop, excessive loading can 3-D mesh and the weather information vided by the NTSB UAS Team could be result in excessive torsion bar deflection, provided by the NTSB meteorologist used effectively in OpenFoam (a widely increasing the likelihood that the geared assigned to the investigation to perform tab linkage could travel to an overcen- used open-source CFD modeling soft- the CFD model analysis. The weather tered position (see Figure 9). ware), more processing was needed. The information included variable wind mag- The CFD results assisted the NTSB specialist combined the mesh files and nitudes, directions, and gusts; pavement systems investigator in designing a more used the open-source software Blender temperature; and additional meteorolog- representative test plan for the airplane’s to manually make the combined mesh ical data. elevator system, more specifically the “watertight” with essentially no holes. The wind simulation revealed that, at interaction with the elevator stops, which The watertight combined mesh file was the accident airplane’s parked location, included a series of static and dynamic further processed using an OpenFoam the elevators could have been subject- elevator load tests to determine what utility to apply settings that allowed for ed to hangar-generated turbulent flow conditions, consistent with the CFD model stability. The final mesh used for that included a small-scale 30+ me- results and the known circumstances the CFD model wind simulation was a ters-per-second (58+ knots) gust mov- of the accident, could enable the geared blend of the combined drone mesh data ing over the airplane. According to the tab linkage to move to an overcentered inserted into a 500 x 500 meter flat plane. simulation, another possibility included position and jam the elevator. (A flat, smooth plane was necessary along small-scale 15/+15 meters-per-second In preparing the test plan, investigators the model boundaries for model stability.) (-29/+29 knots) vertical vector couplets used the hinge moment formula spec- Finally, an MD-83 model was positioned that moved horizontally toward the air- ified in 14 CFR 25.415 to calculate the hinge moments needed to simulate static elevator loads from ground gusts of 25, 55, 60, 65, 70, and 75 knots (the maximum recorded wind gust at YIP was 55 knots). The CFD model results indicated the potential for turbulent gust flow around the airplane’s elevator to induce both lift- ing and downward loads. Thus, for each dynamic test, the load was applied to the elevator, then the elevator was raised to either a neutral or full-trailing edge up (TEU) position (using a forklift and lifting straps) before it was released (using a quick-release mechanism). Releasing the lifted elevator from either the neutral Figure 8. Vertical cross-section visualization of wind simulation results for a discrete or full TEU position allowed it to travel time showing horizontal (“U”) and vertical (“W”) wind magnitudes in area of the accident downward and dynamically contact the airplane’s elevators. trailing edge down stop.

10 • July-September 2019 ISASI Forum None of the static test cases resulted in an overcentered condition of the geared tab linkage. Dynamic tests for simulated gust loads of 25 and 55 knots with the elevator starting in either a neutral or full TEU position did not result in the geared tab linkage becoming overcentered. For the 60-knot simulated gust load, the linkage became overcentered for only the full TEU initial elevator position test. For the 65-, 70-, and 75-knot simulated gust loads, the linkage became overcentered for the neutral initial elevator position tests. Based on the neutral initial position test results, tests of the full TEU initial position were not performed. Figure 9. Elevator stop arm and torsion bar assembly. Conclusions: value of sUAS to derstood, using them to refine drone-cap- depths. Applications (such as this exam- investigation tured imagery was a superior method ple) that allow for a detailed microclimate Increased efficiency in creating a more compared to other, more crude methods. survey (whether affected by a man-made accurate 3-D model The Boeing CFD specialist also believed object, as here, or natural terrain) are only Use of the sUAS to acquire the imagery that drone data mesh products enabled the beginning. A rich 3-D presentation of of the hangar area greatly reduced the the creation of a realistic, high-resolution the accident area, with high geographical time and labor costs required to create a model that would be difficult, if not im- accuracy, allows for further examination 3-D model of the complex building and possible, to create manually with accept- such as geographic information system terrain environment, which would oth- able accuracy. analysis of airport environments, visi- erwise require using blueprints, photo- bility evaluations, and terrain mapping graphs, terrain data (from Google Earth, Value of CFD results in developing elevator and modeling. The wind study enabled the U.S. Geological Survey, and/or other system testing by this exercise is but one of a multitude sources), and CAD software to manually The results of the CFD, including the of ways to think about all the dimensions create the model. For example, it would detailed model of the hangar, greatly of an accident scenario and visualize be difficult to manually create the basic increased the confidence in the wind everything from the overall scene and en- hangar structure alone, but to then add simulation’s ability to resolve the micro- vironment, down to very specific details all of the intricate accessories (duct work, climate environment sufficiently to draw of items. chimneys, ladders, detailed architecture, conclusions about the wind’s effect on Other applications conducted by the the airplane’s elevator. The visualization NTSB UAS Team include viewpoint bushes, trees, terrain near the river, et of the wind patterns in the vicinity of the reconstructions for both aircraft and cetera) could take weeks of labor. Further, tail aided in the development of the test surface vehicles, analysis of terrain at such a monumental effort still might not cases, specifically the dynamic test cases. both large and small scales, and the devel- provide the level of accuracy needed. The results from the dynamic test cases opment of accident site models that During the process, the NTSB UAS helped the investigation understand the enable investigators the ability to contin- Team learned that some additional man- interactions of the elevator stop system ue to take accurate accident site meas- ual processing of the 3-D model created during an impact-loading condition. urements long after the wreckage has by the NTSB UAS Team was required to Without the CFD study information, been cleared. The future of accident prepare the mesh data for effective use by questions could have remained regarding investigation will certainly take increas- the CFD software (such as sealing holes to whether the flow conditions simulated ing advantage of this revolution in create a watertight mesh, which is crucial during the full-scale elevator testing were technology; miniaturized, high-resolution for model stability). This understanding even possible given the known conditions cameras and precise positioning and of the steps involved, gained through on the day of the accident. flight plans that provide data for use in trial and error, will enable more efficient affordable, user-friendly processing preparation and processing of future Future of sUAS use in investigations programs allow investigators to link models. This cooperative and detailed investiga- different facets of their work in new ways. For effective use in the CFD simulation, tion underscored how investigators can The photogrammetry products developed which has 0.5 meters (1.6 feet) resolution, and should explore the confluence of from these efforts provide investigators the input model must have at least that modern technologies in furthering their with the tools to immerse themselves in level of accuracy to model small-scale investigations. The sUAS-based imagery, the accident evidence better than ever turbulent wind patterns caused by the subsequently processed by current gen- before. The sUAS represents far more than obstructions. The Boeing CFD special- eration photogrammetry tools, provided just a means of taking a nice aerial ist believed that, once the processes for investigators the ability to examine the photograph; it is the gateway to deep preparing a watertight mesh were un- environment of an accident scene to great analysis of the accident environment. July-September 2019 ISASI Forum • 11 By James Buse, Senior INVESTIGATING A UNIQUE Technical Lead System Safety, and Jeff Kraus, UAS Flight Test Safety Lead, UAV GEAR COLLAPSE the Boeing Company Photo: Boeing

First flight of the Phantom Eye UAV.

he Phantom Eye is a one-of-a-kind, the planned test points with the exception the stratosphere. Examples include battle- liquid-hydrogen-fueled unmanned of the lakebed landing at Edwards Air field and border observation, port security, air vehicle (UAV) test bed designed Force Base in California, USA. The result- or telecommunications, to name a few. Tto operate at high altitude and is ing unplanned recovery event threatened The Phantom Eye prototype was designed capable of long endurance for persistent the continuance of the flight test program. to stay aloft for five days at an altitude of intelligence, surveillance, and reconnais- This paper presents the challenges of 65,000 feet without landing or refueling. sance and communications missions—an investigating an experimental unmanned Specifications of this prototype include eye in the sky. The demonstrator aircraft demonstrator aircraft, lessons learned, • 150-foot wingspan, can maintain its altitude for multiple days and planning preparations to ensure for a while carrying a 450-pound payload. Typi- successful investigation. • 45-foot length, cal payloads include multiple sensor pack- • Empty weight of approximately 6,100 ages for monitoring, tracking, and commu- pounds, nications. A full-size Phantom Eye variant Description is designed to stay aloft for more than a The Boeing Company’s Phantom Eye is • Takeoff gross weight of approximately week and carry a payload of 2,000 pounds. a high-altitude/long-endurance (HALE) 8,265 pounds, The inaugural flight of the demonstrator UAV. There are various operational needs • Triplex vehicle management system aircraft was marked a success across all for having a long-endurance air vehicle in (VMS) for redundancy, 12 • July-September 2019 ISASI Forum • All-composite structure, valves, an overpressure explosion could occur. A (Adapted with permission from the authors’ technical • Two 2.3L Ford hydrogen engines, great reminder for test personnel approaching and monitoring the state of the air vehicle: “The paper titled Investigating • Three-stage turbochargers, one-stage gear- system is always actively trying to kill you!” a Unique UAV presented during ISASI 2018, Oct. box, variable pitch, three-blade propellers, The team needed to ensure there were no 30–Nov. 1, 2018, in Dubai, • Two eight-foot diameter vacuum dewar leaks of the highly volatile liquid (or gaseous hy- the United Arab Emirates. tanks, and drogen) before approaching the vehicle to start The theme for ISASI 2018 the incident investigation along with the defue- • Liquid hydrogen to gaseous hydrogen heat was “The Future of Aircraft ling and purging process to get into a safe condi- exchanger. Accident Investigation.” The tion. Normal fuel system pressure and expected full presentation can be quantities were verified by the pilot before the found on the ISASI web- First flight timeline launch crew was cleared for approach. site at www.isasi.org in the At 6:21 a.m. on June 1, 2012, the Phantom Eye The recovery crew verified there was no fire Library tab under Technical Presentations.—Editor) performed a 26-minute first flight, up to 4,000 or leakage using an infrared camera, along with feet. The following is a high-level overview of the hydrogen detectors. Once confirmed safe for time: recovery, the mishap plan was initiated at the • 6:21 a.m.—Takeoff from lakebed Runway 15, ground control station with notifications made, with a transition to normal flight guidance. data preserved, and statements collected per the premishap plan, and the UAV was recov- • 6:29 a.m.—All systems and performance ered from the Edwards Air Force Base lakebed normal; pilot commanded the aircraft to runway to the hangar area for defueling and enter the south lakebed holding pattern per purging the fuel system. plan. • 6:36 a.m.—As planned, the pilot manually Investigation commanded gear down. Gear extension The mishap occurred on Friday, June 1, 2012. observed by chase aircraft. Range officer, The aircraft was rendered safe and secured, and James Buse telemetry pilot display all indicated down all evidence was sequestered awaiting arrival and locked gear. of the investigation team to begin the accident • 6:40 a.m.—Cleared to land, investigation process on the afternoon of June 2. The process initiated with a team meeting to • 6:47 a.m.—Landing on lakebed Runway understand eyewitness accounts, read state- 33. Vehicle dynamics were as predicted on ments, discuss the investigative process to final approach. With sink rate on target, the include general plan forward, and conduct a site main skid touched down normally followed visit to capture lakebed witness mark evidence, quickly by nosewheel separating upon including mapping and measuring indentures, contact with the runway and nose strut skids, and the overall footprint. bending back from drag force, with finally The onsite investigation spanned slightly Jeff Kraus main skid collapse. more than one week with many activities occur- ring in parallel. There were three key activities Recovery that contributed to finding the root cause. In no video—a trifecta of data that After the landing gear collapsed, the air vehicle particular order of priority, they were photo- complimented and confirmed was in an unknown state. The team needed to grammetry evidence compilation and analysis, the events as they unfolded that use the utmost caution to approach and place design pedigree research, and engineering Friday morning. the vehicle into a safe state. The telemetry investigations (EIs). The onsite aspect of the Alignment and agreement of showed no issues or anomalies with the fuel latter consisted of identifying components and the data were comforting in the system detected up until the vehicle shutdown. arranging logistics for follow-on detailed EIs in final analysis, but the investi- The tank active vent behavior and pressure rise the St. Louis, Missouri, metallurgical lab. gation was accelerated in the rate following the landing were as expected— initial days with the benefit of once again, up until power off from the vehicle Photogrammetry long-range video recorded in shutdown. Test personnel approached with The fortunate aspect of conducting flight high fidelity. The video evidence hydrogen detectors that indicated no hydrogen testing on a range is that you have the luxury of permitted review to be slowed leakage. operating in a controlled environment and are to a frame-by-frame basis with- Anything filled with liquid hydrogen should afforded the opportunity to build contingency out introducing the negative capture and keep your attention as long as it is management and data acquisition into the plan. pixelating that often hampers fueled—even while the engines are off and the Data acquisition served the program and the video quality. This clarity vehicle has no power, in what appears to the investigation quite well. Telemetry data, captur- allowed photogrammetry uninformed observer in a calm state. It is quite ing vehicle performance parameters, became measurements and assessment dangerous. Pressure is constantly building as instrumental in aligning witness marks with to be performed, narrowing the liquid hydrogen warms and turns to gas, the vehicle environmental recovery data, which the window of speculation as constantly increasing the pressure inside the corroborated the evidence captured through to the mishap’s root cause. The tanks. If the valves stick or moisture freezes the the lens of long-range, onboard, and chase significant events that occurred July-September 2019 ISASI Forum • 13 during the lakebed recovery were easily of dynamic load modeling and testing. instability of liquid hydrogen required identified and quickly steered the investi- Interesting to note that the dynamic load the team to defuel the aircraft and gation focus to further delve into the nose modeling and testing performed purge the lines. As it turned, it was not landing gear structure, which unveiled postanomaly predicted these failure a factor; however, variables such as fuel further evidence leading to root cause modes. 2) The piston “as built” configura- contamination or exact weight at land- identification. tion did not conform to the “as designed” ing could have been lost critical data in version, which was due to a failure in the the pursuit of getting to root cause. Design pedigree nonconformance review process. Research into the history of the nose land- No pilot first account ing gear structure became a primary focus. Recommended actions Even with onboard cameras, chase, and The examination centered on reviewing There were three recommended actions long-range cameras, there is no “in the the evolution of the design—specifically that came as a product of the mishap seat” or “feel” that an onboard pilot may the latest or current design version com- investigation and root cause corrective have during an event. With no voice re- pared to the installed version. The drawing action. The first two were engineering corder on board, sounds from the event revealed an engineering change order in- design, and the third was engineering pro- cannot be detailed as with a pilot or itiated circa 1985 stipulating removal of a cess. First, it was recommended that the cockpit voice recorder (CVR). The team counter bore. This counter bore proved to nose landing gear be redesigned and that is limited to onboard telemetry and cam- be in the origin of the gear failure location. its performance be verified using dynamic eras being functional through the event. load modeling and testing. Second, the Telemetry Limitations: In addition to Investigation main landing gear was further scrutinized no CVR, the vehicle configuration did The EIs consisted of visual and magnified via analysis and testing to ensure that not include an onboard crash-surviv- fracture surface inspection and material adequate landing safety margins existed. able data recorder. The team did have pedigree review. All visual inspections Third, conduct a success preliminary and live-streamed telemetry data, but it was revealed classic ductile overload fracture. critical design review on the redesigned limited to certain parameters. However, The fracture origin was determined to nose landing gear and main landing gear the team had accurate indications of be on the forward side or zero-degree to include independent subject-matter the fuel state along with gear position position of the outer fillet radius between experts. (down and locked). Had the telemetry the lower cylinder and the larger diameter stream been disrupted, those indica- upper cylinder. The piston microstructure Lesson learned tions/confirmations would have been was normal with no internal defects. No A lesson learned was test only what you in- unknown, adding to the complexity of other anomalies were noted at the fracture tend to test. In other words, independently the investigation. origin or other elements of the fracture verify that each item of the configuration zone. The overload fracture propagated build has the demonstrated credentials to Key preparation ingredients for a upward through the thickness and out- safely support the future test objectives. successful investigation ward circumferentially in both directions Trust your baseline engineering data but • Controlled test range around the cylinder. Aft bending load from verify the data to be accurate, current, and • Chase aircraft landing drag forces tore the piston lower representative of the configuration to be cylinder with the attached wheel assem- tested. • Video (long-range, short-range, air- bly out of the upper piston cylinder. This borne, stills, onboard, ground chase, action resulted in the remaining landing Challenges and GoPro™) gear stub ultimately collapsing under the Vehicle Peculiarities: Investigating a mishap • Telemetry balance of remaining frictional forces. of a one-of-a-kind aircraft has a unique set • Lakebed (optional—can convenient- of challenges. Although this air vehicle was ly capture witness marks) Investigation summary based on a previous air vehicle and mul- • Flight test venue limited exposure to The program responded to the first flight tiple baseline contributors were designed competing users/traffic anomaly by initiating an accident investi- and documented in the 1980s, as a result gation team and a root cause corrective ac- of an engineering hiatus there was a lack tion team. Both were formed to investigate of data for reference due in part to record Summary the incident, determine proximate cause, retention policies. The potential to review In recognition, the test team successfully and recommend corrective actions. The past test reports for similar landing gear navigated a high-stress, off-nominal first nose landing gear experienced a greater events or issues was nonexistent. Unlike flight event in which the vehicle was than expected vertical load, but much investigating a production aircraft mishap recovered safely. The vehicle was repaired lower than the maximum specified load in which a trove of previous mishap data and updated with a redesign of the that resulted in a piston shaft failure. The exists, a one-off air vehicle will have lim- landing gear and flown eight more times, greater than expected load was caused by ited or no production data or history to achieving greater altitudes and endur- a combination of unexpected bearing fric- reference. ance times with each subsequent flight. tion, piston shaft bending, and increased Fuel Instability: The unstable nature of the Lessons learned were archived for a drag on the main landing gear skid. These hydrogen fuel introduced an additional follow-on version along with passing on unexpected behaviors of the piston where investigation challenge—which comes additional best practices to future UAV primarily due to two factors. 1) The lack in the form of evidence preservation. The test teams.

14 • July-September 2019 ISASI Forum (Adapted with permission from the author’s technical paper titled Application of Innovative Investigation Technologies presented during ISASI 2018, Oct. 30–Nov. 1, 2018, in Dubai, the United Arab Emirates. The author noted that together with his col- league Albert Urdiroz, he has been involved in the investigation of an A380 that suffered failure on engine No. 4 while in cruise over Greenland on Sept. 30, 2017, which forms the basis of this technical paper. The theme Application of Innovative for ISASI 2018 was “The Future of Aircraft Accident Investigation.” The full presentation can be found on the ISASI website at www. Investigation Technologies isasi.org in the Library tab under Technical Presentations.—Editor) By Sundeep Gupta, Accident Investigator, Product Safety, Airbus Event synopsis On Sept. 30, 2017, an A380 fitted with Engine Alliance GP7200 engines suffered an uncontained engine failure while cruising at 37,000 feet over Greenland. Engine parts were liberated onto the ice sheet below, including the fan hub, fan blades, and fan casing. The aircraft diverted uneventfully to Goose Bay, Newfoundland, Canada. Retrieval of the engine components was crucial to understanding the root cause. The key component of interest was the fan hub, a 250 kilogram piece of titanium 80 centimeters in diameter. The fan hub is the central rotating component to which the fan blades are attached. The investigation team faced a huge challenge due to the geography and the extreme climate and looked to technology to help search for the fan hub. Figure 1. Initial satellite imagery. The Annex 13 investigation was delegated by the AIB of Denmark to the French BEA, assisted by accredited representatives and advisors. Airbus, as nominated advisors, deployed resources and technologies to support the investigation and the search for the liberated parts on the Greenland ice sheet.

Innovative investigation technologies Satellite imagery To visually identify the location of any parts, Airbus immediately launched satellite imagery of a 20 kilometers x 20 kilometers area using the Pleiades satellite constellation, owned and operated by Airbus De- fence and Space. The two satellites are in continual orbit around Earth and can be programmed to take images at a resolution of 50 centime- ters of any given area in approximately six hours. The first images received were obscured due to cloud cover (see Figure 2. Flags on the Greenland ice sheet. Figure 1). The satellite continued to take images daily; and over the next few days, the cloud disappeared and we received the first images of the visible ground beneath. However, a blanket of snow had fallen, making it impossible to visually detect any engine parts. Greenland has a climate in which the snow never melts to its previous level. Only a percentage of the annual snowfall melts each year, meaning any parts covered by snow will never surface again. While satellite imagery did not provide any tangible results during this investigation, it may prove useful for other events such as locating an accident site in remote or inaccessible areas, mapping large accident sites, or runway excursions. Figure 2 shows an image of flags on the Greenland ice sheet, obtained later in the investigation, that were possi- ble investigation sites captured by the satellite images.

Ballistic analysis The potential search zone covers hundreds of square kilometers. To Figure 3. Ballistic analysis of engine fragments helped nar- narrow the search area, Airbus advisors worked with Ariane Group spe- row the search area.

July-September 2019 ISASI Forum • 15 cialists to quantify the most likely trajectory of the components from their release at 37,000 feet to landing on the ice sheet. Examination of the engine revealed that the fan hub had ejected in several pieces. Ariane Group’s analytical models (see Figure 3, page 15) were able to narrow the primary search area to 2 kilometers x 4 kilo- meters when provided with properties of a fragment, such as altitude, aircraft speed, wind, etc. The ongoing investigation established the fragment ejection angle and ejection speed (see Figure 4). This allowed the area to be refined Figure 4. Parameter effects of the liberated parts. to 1 kilometer x 2 kilometers. The parameters of the liberated parts influenced the ballistic analysis. Smaller, heavier parts would be pro- jected forward of the event point while lighter, larger parts that have aerodynamic drag would be carried rearward of the event point in the direction of the wind. The fan hub primary search area shown in Figure 4 was identified us- ing both Ariane and U.S. National Transportation Safety Board (NTSB) ballistic analyses, which were largely consistent. Given the extreme climatic conditions, searching the identified area with ground teams remained a challenge. The investigation team therefore tried to locate the fan hub using airborne radar scanning.

Synthetic aperture radar scanning Figure 5. Synthetic aperture radar scanners locate metal Synthetic aperture radar is normally used for geographical surveys. objects buried in the snow This was the first time it would be used to detect metallic objects bur- ied under snow. ONERA, a French aerospace lab, provided two radar pods equipped with three radars, each scanning at a different bandwidth. These were mounted on a Falcon 20 aircraft provided by AVDEF, part of the Airbus Defence and Space portfolio (see Figure 5). The team assembled in Greenland and first performed a trial to verify equipment functionality and to calibrate the data to the GPS position. A calibration test piece was mounted on a tripod and positioned on a snow-covered golf course with its exact angle and GPS position noted (see Figure 6). The aircraft overflew the area at the two radar operating Figure 6. A calibration test piece is mounted on a tripod and altitudes to capture the landscape with all three radar bands. positioned on a snow-covered golf course with its exact Once the image was processed, the radar scan revealed the ground angle and GPS position noted. beneath the snow and ice (see Figure 7). Examination of the image confirmed that the test piece had been de- tected, indicated by a bright white return on the radar image. A second radar return was seen on the image in the vicinity of the test piece. The cause of this return was not known and when the team returned to the site, metallic fencing wire buried under the snow was discovered (see Figure 8). This demonstrated that the technology was fundamentally capable of detecting metallic objects even when buried under snow. Figure 7. A processed image shows the ground and detects An extensive area for the hub search was scanned and data was the test piece. collected. However, crevasses were present within the area and created background noise in the radar data. The radars were able to scan about 36 meters below the surface of the ice. Figure 9 shows an area approximately 4 kilometers x 2 kilometers. Each of the 200 million pixels, each covering 20 centimeters2 in X-Band, was scanned by a total of 72 images at different angles and polariza- tions. The large amount of data underwent complex posttreatment to differentiate crevasse returns and background noise from credible fan hub targets.

3-D laser scanning Figure 8. The test reveals a metal fence buried beneath the Shortly after the event, the BEA and AIB of Denmark scrambled heli- snow. copters along the aircraft track before more snow fell and were able to

16 • July-September 2019 ISASI Forum Figure 10. The BEA and AIB of Denmark use helicopters to Figure 9. This image shows an area approximately 4 kilometers x 2 locate and recover the larger, aerodynamic engine compo- kilometers. nents. locate and recover the larger, aerodynamic engine components that had been liberated during the event. Due to their size, they were easier to spot by the helicopter crew (see Figure 10). With their aerodynamic properties, these components were blown rearward of the event pointing in the direction of the wind, whereas the fan hub trajectory was analyzed to have landed in front of the event point. Because the fan hub had not been located, the investigation team needed to ensure that maximum data was extracted from these recov- ered parts. The data would also support any hypothesis put forward and validate any analysis. Airbus deployed the use of state-of the-art 3-D scanning technology provided by IDLAB, a subsidiary of Airbus commercial (see Figure 11). Figure 11. Airbus uses state-of the-art 3-D scanning technol- Using 3-D laser scanning equipment, all of the parts were digitized, ogy to create 3-D models of the retrieved components. creating 3-D models of the retrieved components and capturing details with an accuracy of 0.03 millimeters. The scanned data allowed an assessment of which parts had been recovered and which remained on the ice sheet. A 3-D reconstruction was produced, which provided insight as to how the engine event may have unfolded (see Figure 12). In addition, 3-D models can be overlaid to reference data such as Catia models or scans of reference parts to analyze how a part became distorted or deformed (see Figure 13). 3-D scanning can provide an accurate record of parts as they are recovered and before any disassembly or destructive testing takes place. The data can be made available to all parties of the investigation so that analysis can begin simultaneously. The data can be imported into Catia or viewed as a 3-D pdf model. Any hypothesis can be cross- checked against this 3-D data. Figure 12. A 3-D reconstruction was produced, providing insight as to how the engine event may have occurred. Summary The investigation team led by the BEA has been working hard to move the investigation forward. The engineering analysis to date has allowed mitigating actions to be taken on the A380 fleet powered by Engine Alliance GP7200 engines. The investigation, search, and analysis con- tinue. When Airbus is engaged in an investigation, along with Airbus commercial, we can also engage the resources and expertise of Airbus Helicopters and Airbus Defence and Space domains. By utilizing the resources from across the Airbus Group, Airbus advisors have been able to deploy a number of innovative technologies to gain the maximum amount of knowledge possible from the compo- nents already retrieved and to locate the fan hub, the key component of Figure 13: 3-D models can analyze how a part becomes the investigation. distorted or deformed.

July-September 2019 ISASI Forum • 17 5 MILLIMETER CRACK LEADING TO AN ENGINE FIRE By David Lim, Principal Investigator, Transport Safety Investigation Bureau, Singapore

Synopsis countered weather that required them to on the situation, the technical services On June 27, 2016, a B-777-300ER depart- perform weather-avoidance maneuvers. personnel believed that it was a faulty oil ed Singapore for Milan, Italy. About two About 30 minutes into the flight, when quantity indication. As the aircraft had hours into the flight, the right engine the aircraft had climbed to 30,000 feet, the just departed and was not far from Sin- indication showed low oil quantity. Subse- flight crew noticed that the oil quantity gapore, the technical services personnel quently, the flight crew felt a vibration in parameter in the engine indicating and recommended that the aircraft return. the control column and cockpit floor and crew alerting system (EICAS) showed 17 According to the flightcrew members, decided to return to Singapore. units for the left engine but only one unit as they were passing waypoint VPG at Shortly after landing at Changi Air- for the right engine. The flight crew also approximately 3:20 p.m., the first of- port, a fire occurred in the vicinity of the noticed from the EICAS display that the ficer performed a routine fuel quantity aircraft’s right engine. After the aircraft right engine oil pressure was fluctuating check. After comparing the totalizer fuel came to a stop on the runway, a fire de- between 65 and 70 pounds per square quantity with the planned fuel remaining veloped under the right wing. The airport inch (psi) while the oil temperature for quantity, it was determined that the fuel rescue and firefighting (ARFF) service -ex the right engine was 10 degrees Celsius consumption was better than expected as tinguished the fire. Damage to the aircraft higher than the left engine. However, both the fuel on board was 600 kilograms more included heat damage to the core of the the oil pressure and temperature param- than the expected value. engine, portions of the engine cowlings, eters were within the normal operating At 3:28 p.m., the engineer sent a mes- and the wing area directly behind and range. sage via the aircraft communication and outboard of the right-hand engine. The flight crew looked through the reporting system (ACARS) to the flight The Transport Safety Investigation available manuals but was unable to find crew relaying the recommendation of the Bureau conducted the investigation into an appropriate procedure that addressed technical services personnel for the air- this occurrence with the assistance of the low engine oil quantity situation. craft to return to Singapore and request- the U.S. National Transportation Safety At 3:04, the pilot-in-command (PIC) ing the crew to contact the engineering Board, the U.S. Federal Aviation Adminis- contacted the company’s engineering control center. tration (FAA), and the aircraft and engine control center for assistance via satellite The PIC contacted the engineering con- manufacturers. communication. The PIC informed the trol center, and a conference call among The information in this paper engineer on duty of the engine parame- the PIC, the engineering control center, details the findings and other aviation ters and asked if it was safe to continue and the technical services personnel was safety aspects deliberated during this the flight. The engineer informed the PIC held, which lasted about 20 minutes. investigation. that since the oil pressure was within The PIC noted that the flight crew had the normal operating range, there could been monitoring the right engine oil Background of Flight be a faulty oil quantity indication. The parameters for 50 minutes; other than the A B-777-300ER departed Singapore’s engineer advised the PIC to continue with indicated low oil quantity, the parameters Changi Airport at 2:24 p.m. for Milan on the flight but monitor the right engine appeared normal. It was jointly assessed June 27, 2016. The aircraft carried two sets oil parameters. The engineer told the PIC that the flight could continue to Milan of operating crew—four pilots in total. that he would also contact the company’s with the proviso that the flight crew mon- As the aircraft was climbing to its cruis- technical services personnel for advice. itor the right engine oil parameters and ing altitude, the flightcrew members en- After being briefed by the engineer contact the engineering control center for 18 • July-September 2019 ISASI Forum assistance if needed. center, and the technical services per- (Adapted with permission from the author’s Shortly after the conference call ended, sonnel, it was assessed that there was no technical paper titled 5 mm Crack Leading the flight crew felt an unusual vibration need to shut down the right engine and to an Engine Fire—Lessons Learned in the control column and cockpit floor. decided that the aircraft would return to presented during ISASI 2018, Oct. 30–Nov. 1, 2018, in Dubai, the United Arab Emirates. The crew tried to diagnose the problem Singapore with the right engine operating The theme for ISASI 2018 was “The Future by changing the engine power settings at idle power. In the midst of the confer- of Aircraft Accident Investigation.” The full and found that the vibration disappeared ence call, the in-flight supervisor (IFS) presentation can be found on the ISASI when the power of the right engine was informed the flight crew that there was website at www.isasi.org in the Library tab reduced. At about the same time, the a burnt smell detected in the cabin. In under Technical Presentations.—Editor) flight crew caught a momentary wisp of a response, the flight crew turned off the burnt smell in the cockpit, but the smell right engine bleed system. disappeared quickly. According to the cabin crew, the smell At 4:04 p.m., the PIC informed the was particularly strong in the busi- engineering control center about the vi- ness-class cabin, in the forward part of brations experienced whenever the right the aircraft. The cabin crew distributed engine was operated at higher power wet towels for the passengers to hold over settings. In the ensuing conference call their nose and breathe through. among the PIC, the engineering control After the conference call ended, the

Party Time Speaking Content

6:51:50 p.m. PIC How is it looking? Is the fire contained?

We are still trying to contain the fire.… The fire is 6:51:53 p.m. FC pretty big.… Will like to advise…disembarkation on your port side.

6:52:05 p.m. PIC Okay, evacuate from the port side confirm. Still trying to contain the fire.… Still some random fire 6:52:09 p.m. FC on your right-hand engine, but we are keeping it under control. 6:52:24 p.m. PIC Do you need us to evacuate from the port side? 6:52:29 p.m. FC Singapore 368 standby, standby.

6:52:33 p.m. PIC Okay standby for your instructions Singapore 368. Standby for your instructions. David Lim July-September 2019 ISASI Forum • 19 Thus, the flight crew performed its Time Party Content own fuel calculation by subtracting the Speaking amount of fuel consumed from the total amount of fuel at the start of the flight. 6:54:08 p.m. FC We have kept the fire under control. We will like to The fuel consumed was calculated by advise disembarkation on your port side. multiplying an average fuel flow value (that the flight crew determined) by the Okay, you want us to disembark through the slides or duration of the flight. The crew arrived at 6:54:20 p.m. PIC are you going to provide mobile stairs? a figure of about 79 tonnes. As this tallied well with the totalizer fuel quantity figure, We will like to advise disembarkation on your port the flight crew concluded that the fuel 6:54:38 p.m. FC side. disagree message was a spurious one and Okay, you want us to disembark on the port side that there was no need to proceed with 6:54:48 p.m. PIC through the emergency slides. Can you confirm that? the fuel leak checklist. Several times on the return journey and 6:55:14 p.m. PIC Can you just confirm that we need to evacuate through as the aircraft approached Singapore, the the left through the emergency slides? flight crew was queried by air traffic -con Negative, negative, negative. We will like to advise trol (ATC) if it needed any assistance. The 6:55:33 p.m. FC disembarkation, disembarkation. No evacuation, no flight crew replied that, other than the evacuation. need to fly at the lower altitude of 17,000 feet, no assistance was needed as all other 6:55:42 p.m. PIC Okay, disembarkation through mobile steps under- stand, understand. operations were normal. Prior to landing, the flight crew jetti- flight crew reduced the right engine to is when the totalizer fuel quantity is less soned about 41,500 kilograms of fuel to idle power and proceeded to turn the than the calculated fuel quantity. bring the aircraft to below its maximum aircraft around to return to Singapore. The flight crew observed from the landing weight. For the return journey, the flight crew display of the flight management system At 6:49 p.m., the aircraft landed on the adopted the procedure for single-engine that totalizer fuel quantity was about 79 runway. About 20 seconds after the thrust operation, including a descent to 17,000 tonnes, and the calculated fuel quantity reversers on both engines were deployed, feet before reducing the right engine to was about 83 tonnes. the occupants in the cabin heard two idle power. However, the flight crew did not per- loud bangs, accompanied by two flashes, When the IFS informed the flight crew form the fuel leak checklist. According originating from the right engine area. that the burnt smell in the cabin was still to the flightcrew members, they believed At the same time, the flight crew heard a present, the right air conditioning pack the calculated fuel quantity was no longer soft thud. The ARFF personnel who were and recirculating fans were switched accurate in view of the following: monitoring the aircraft’s arrival informed off. Shortly after, the smell in the cabin a. Input changes had been made to the the control tower about the fire at the subsided. flight management system after the right engine. The control tower informed At 5:21 p.m., the flight crew received a right engine was set to idle power. the flight crew of the fire and instructed the aircraft to stop at the intersection be- fuel disagree message on the EICAS. The b. They were no longer on the planned tween the runway and rapid-exit Taxiway flight crew performed the fuel disagree flight route. checklist, which suggested four scenarios E7. The flight crew did not receive any fire in which a fuel leak should be suspected c. They had 600 kilograms more fuel warning in the cockpit. and when the flight crew should perform than expected when they last per- the fuel leak checklist. One such scenario formed a routine fuel check. Response to fire ARFF, which was on standby with four foam tenders and one water tender, en- tered the runway as soon as clearance to enter was given by the control tower. The first foam tender arrived on scene after 57 seconds and started discharging foam at the right engine. When responding to the fire, the fire commander (FC) of the ARFF requested the control tower to ask the flight crew to switch the aircraft radio to the emergency channel for communication between the Fuel flows in the tubes while oil flows in the FC and flight crew. cavity around the tubes for heat exchange. Subsequently, the FC and PIC estab- lished communication on the emergency Figure 1. Schematic diagram of the main fuel oil heat exchanger. 20 • July-September 2019 ISASI Forum by both the engine fuel and oil system, was exam- ined for the presence of an internal leak. The MFOHE contains a series of tubes. Fuel flows in these tubes, while oil used for lubricating the engine flows around the tubes (see Figure 1). This allows oil to be cooled through heat transfer to the fuel through the tubes. The design of the MFOHE is such that the oil and fuel flow paths will not cross, and the oil and fuel will not come into contact with each other. The MFOHE was removed from the engine, and the preliminary pressure tests performed on it confirmed an internal leak between the oil and fuel flow paths. The MFOHE was sent to the engine manufacturer’s facility where a computer tomography scan was performed. The scan results showed that there was a cracked fuel tube that was displaced. The MFOHE was then sent to the manufacturer’s facility for further examination. In a test performed to simulate the operation of the MFOHE at idle engine power setting, the leak rate from the displaced cracked tube was found to be about 31 pounds per minute. A portion of the MFOHE casing was removed. One of the Figure 2. View of cracked tube in the main fuel oil heat exchanger. fuel flow tubes was found cracked and displaced (see Figure 2). 1. The cracking of a tube in the MFOHE allowed fuel channel, and key exchanges occured (see page 19). in the fuel flow path of the MFOHE to flow into the Jet fuel that was discharged from the right engine onto the tarmac oil flow path in the MFOHE. The investigation has fueled a fire that impinged on the underside of the right wing near the not revealed other sources of fuel leak. right engine area. ARFF managed to bring the fire under control and put out the visible 2. During all phases of the engine fuel pump opera- fire in the right engine area and on the ground at 6:53. However, ARFF tion, fuel is delivered at pressures between 400 and personnel, using an infrared detector, found a heat signature within 1600 psi. In comparison, the pressure within the the internal section of the engine and continued to monitor the engine oil system is about 100 psi. As such, when situation. Key exchanges between the FC and PIC at that point occured the fuel carrying tube in the MFOHE cracked, the (see page 20). higher pressure fuel entered the engine oil distribu- About three minutes later, the fire appeared again at the forward sec- tion system. tion of the right engine. It was immediately put out by ARFF. There was 3. During the normal operation of the engine oil no fire warning in the cockpit when the flight crew was informed of the system, a small amount of oil will collect in the ongoing fire by the FC. The flightcrew members eventually discharged A sump. However, when fuel leaked into the oil both the bottles of fire extinguishing agent into the right engine when system, it filled the A sump until its maximum they were queried by the FC if they had discharged the bottles. Eventual- storage capacity. The additional quantity of leaked ly, after the FC had assessed the situation to be safe and that the fire was fuel overflowed into the booster spool cavity and brought under control, the occupants of the aircraft disembarked via started to collect there (see Figure 3, see page 22). mobile stairs. Once the booster spool cavity was filled to the aft lip, the excess fuel leaked through a gap between the spool Damage to aircraft and engine and aft stage booster vane into these areas (see Figure The right wing and engine area of the aircraft sustained extensive dam- 4, see page 23): age. The most extensive damage to the right wing was in the vicinity of • HPC through the core airflow. the right engine. Fire damage was also observed on the underside of the • Fan duct when the VBV doors are open at engine right wing, outboard of the engine. The fan section, variable bleed valves idle power. (VBVs), and high-pressure compressor (HPC) of the right engine sus- tained varying degrees of heat damage. The oil tank and various engine drain points are areas in which one would usually expect to find only oil. Instead, fuel was found in those locations. Similarly, Cause of fuel leak residual fuel was found in the various engine sumps. In During the initial postoccurrence examination of the right engine, fuel addition, the gearboxes and the engine bearings, which was found in the booster spool cavity, oil tank, all bearing sumps, and are usually coated in oil, were dry. These observations accessory and transfer gearboxes. These were areas of the engine where suggest that engine oil was displaced from the engine fuel should not be present during normal operation. and fuel, in place of oil, was distributed throughout the The main fuel oil heat exchanger (MFOHE), a component that is used engine oil system. July-September 2019 ISASI Forum • 21 Figure 3: Process of fuel filling the A sump and booster spool cavity.

Engine oil lubricates and cools the tion. This was consistent with the flight landing. engine bearings and gearboxes and helps crew’s observation that the vibration There was no fire during the airborne to lower vibration at the engine bearings. seemed to disappear when the engine segment of the aircraft’s return journey Fuel is not as efficient as oil for engine was at reduced power setting. to Singapore. This was due to the high lubrication. Therefore, when oil had For the remainder of the return journey velocity of the airflow over the exterior been displaced by fuel in the occurrence back to Singapore, fuel leaked through of the engine, which prevented both the engine, oil temperature increased. The the core of the engine and the fan duct. As ignition and sustained combustion of the temperature increase was a result of fuel the engine was operating at idle power, leaked fuel. in the oil system that was not able to cool the VBVs were open, allowing the leaked As the aircraft arrived to land, fuel was the engine bearings and gearboxes as fuel into the VBV ducts and the fan duct, still leaking from the engine through var- efficiently as oil. where it could accumulate in the honey- ious leakage areas (see Figure 4). When The vibration detected by the flight comb core material behind the perforated the thrust reversers were deployed, the crew when operating the right engine at a walls of the thrust reverser duct. velocity of airflow over the core exhaust higher power setting was likely due to the nozzle was significantly reduced. The area fuel that collected in the booster spool aft of the turkey feather seal, which is a cavity. This cavity is a dome-shaped space, Fire initiation and propagation protrusion on the core exhaust nozzle, and rotational forces would have caused The investigation team determined that would have experienced the most signifi- the fuel to be spun against the inner wall the fire was a result of hot surface ignition cant disruption of airflow. In addition, the of the booster spool cavity as the engine of leaked fuel at the area behind the accumulated fuel in the fan duct was also was operating. The rotating fuel created turkey feather seal of the core exhaust distributed over a wide area of the lower imbalance that resulted in vibration. At nozzle. Based on recorded video and data surface of the wing. higher engine power settings, the vibra- from the aircraft, the investigation team The investigation team believed that, tion would have been more pronounced believed that the fire first started after the with the disrupted airflow, the mixture as compared to the engine at idle opera- thrust reversers were deployed during the of accumulated fuel on the core exhaust 22 • July-September 2019 ISASI Forum in the same location as the previous two units examined. In August 2014, a B-777-300ER installed with GE90-115B engines and operated by another operator experienced after landing and engine shutdown a small, candle-wicking-like fire emanating from its left engine center vent tube. Teardown of the MFOHE revealed that fuel had entered the oil system through a cracked tube. This MFOHE was manufactured be- fore the improved manufacturing process was implemented. The cause of the crack in the August 2014 event was determined by the engine and MFOHE manufacturers to be the Figure 4: Leak path for fuel. variation in the crimp on the tube that resulted in contact between the support nozzle and fuel in the airflow would have History of MFOHE leakage plate and crimped tube. The contact re- been sufficiently heated to the point of According to the MFOHE manufacturer, sulted in stress concentration that could ignition. Subsequently, the fire propagat- prior to December 2013, there had been have led to crack initiation. ed through these following areas of the nine instances of leaking MFOHEs on The August 2014 occurrence led the engine: GE90-115B engines that were returned to engine manufacturer to introduce a the MFOHE manufacturer for repair. The diagnostics program to monitor oil • fan duct, causes of the leakages in these nine MFO- consumption trends. After an aircraft has • thrust reverser blocker doors, HEs, which would have required destruc- landed, the aircraft operator will send the • booster, tive examination, were not determined. engine data related to the preceding flight Between December 2013 and February to the engine manufacturer for analysis • HPC, and 2014, the MFOHE manufacturer received by the diagnostic program. Should the • VBV. three MFOHEs from GE90-115B engines diagnostics program detect any abnor- As the engine was spooling down, the that were suspected to be leaking. The mal oil consumption trend related to a excess fuel that had been collected in engine and MFOHE manufacturers jointly suspected fuel leakage into the oil system, the booster spool cavity was discharged decided to conduct destructive exami- the operator will be alerted by the engine nation of these three units. It was found through the fan duct and flowed onto the manufacturer. that two MFOHEs had a partially cracked runway and caught fire. The fuel that was In a failure analysis test conducted tube. The cracks were at the crimped distributed over a wide area of the lower by the engine manufacturer in Septem- areas of the tubes. At that point, the surface of the wing also caught fire. ber 2016, it was further discovered that cracks were attributed to the stress con- unintended diffusion bonding occurred centrations created at the support plate during the manufacturing process of the Design of MFOHE hole edges resulting from the crimping MFOHE when elevated heat was applied. operation. The third unit had a tube with It was identified that the diffusion bond- The MFOHE was designed and manufac- a pinhole leak. ing occurred at the areas where there was tured by a component manufacturer that In April 2014, the engine and MFOHE close contact between the tubes and the supplied it to the engine manufacturer. It manufacturers conducted a review of the support plates. was designed to have unlimited service manufacturing operations. Improvements During normal operation of the lifespan, i.e., periodic replacements would to the manufacturing process were made MFOHE, stress was introduced at the not be needed. The engine manufacturer in May 2014. The improvements included fused area that ultimately led to the tube did not require any periodic inspection of using a standardized crimping tool to cracking. It was also determined that the internal portion of the MFOHE during eliminate the variation in the crimps due crimping increased the likelihood and its service lifespan. to the use of hand tools. In addition, the severity of diffusion bonding to occur. The manufacturing process involves crimped fuel tubes that have a history of crimping, a process in which a force is cracking will be welded close at assembly. applied to achieve a slight deformation The MFOHE manufacturer received Resolution for cracked tube problem of a material. The purpose of crimping another MFOHE unit from a GE90-115B As part of its airworthiness control is to deform the cross-sectional shape of engine suspected to be leaking that system, the FAA, the regulatory authority the tubes so that they cannot slide freely was removed from service in June 2014. for U.S. aeronautical products, requires through the round holes of the support This unit was manufactured before the engine manufacturers to identify plates. This prevents the support plates improved manufacturing process was unsafe conditions and implement from moving during assembly (see Figure implemented. Destructive examination corrective actions. 5, see page 24). revealed a partially cracked crimped tube, The FAA offered a process known as July-September 2019 ISASI Forum • 23 continued airworthiness assessment methodologies (CAAM) to help engine manufacturers identify potential unsafe conditions associated with their prod- ucts. CAAM also helped engine manufac- turers determine if the potential unsafe conditions were likely to exist or develop in other products of the same type design. The engine manufacturers may use CAAM to • assess the risk associated with the unsafe conditions. • develop and prioritize appropriate corrective actions to address the unsafe conditions. • assess the effectiveness of the correc- tive actions. Under the CAAM process, each occur- rence would be accorded a severity level. The severity ranges from Level 1 (minor) to Level 5 (catastrophic). The determina- tion of the CAAM level was based on the actual damage and consequences in the occurrence. The CAAM level of a possi- ble future occurrence was then assessed and used to determine the urgency to Figure 5. Crimps on tubes. implement corrective actions. The unsafe with the SB was brought forward. The additional costs and operational con- conditions identified and the corrective actions called for by the SB were straints for stakeholders in the aviation actions determined had to be approved performed on all affected MFOHEs by industry. In such situations, civil aviation by FAA before being implemented. July 2017. authorities will have to balance the needs The small candle-wicking fire occur- of promoting growth of the aviation rence in August 2014 was categorized industry and to ensure the safety of the as a CAAM Level 2 event by the engine Timeliness of safety improvement flying public. manufacturer. implementation Following its investigation into the At the time of the occurrence, the actions August 2014 event, the engine manufac- called for by SB 79-0034 had not yet been Decision-making during abnormal turer issued Service bulletin (SB) 79-0034 performed on the occurrence engine. situation in December 2014 to address the issue of The engine had last undergone an engine In the initial communication, the FC fuel leakage into the oil system. The SB shop maintenance in March 2014, before advised the PIC, “We are still trying to required the MFOHE to be removed from the SB was issued. contain the fire.… The fire is pretty big.… the engine no later than the next occasion Had the SB been incorporated in the Will like to advise…disembarkation on when the engine was in an engine shop occurrence MFOHE, the fuel leak would your port side.” As the commander of the for engine shop maintenance. The dead- not have occurred and the fire event aircraft, the PIC was aware that the deci- line for compliance with the SB was set would have been avoided. sion to evacuate lay with him and that he in accordance with the CAAM consistent Despite the engine manufacturer could order an evacuation even if the FC with a Level 2 criticality. following the CAAM, which was provided advised a disembarkation. Although the As required by SB 79-0034, the MFOHEs by the FAA, a fuel leak in the MFOHE due PIC was the only person actively commu- were to be sent back to the MFOHE man- to a cracked tube recurred and resulted in nicating with the FC, the other flightcrew ufacturer to check for leakages. Should a more severe consequence of an uncon- members were listening to the communi- a leakage be detected, the openings at trolled fire. cation and the decision not to evacuate the entrance and exit of the leaking tube This occurrence suggests that there was reached collectively. would be welded close to prevent fuel is room for civil aviation authorities to Making a decision to evacuate is not from flowing into it. Crimped tubes that review and enhance, if necessary, their always straightforward: had a history of cracking will also be control system to ensure that correc- • On the one hand, the operator’s flight welded close, regardless of whether the tive actions can be implemented more crew training manual recommends MFOHE was found leaking. expeditiously to prevent the recurrence of that in a situation in which a persis- In response to an immediate safety unsafe conditions. tent smoke or a fire cannot positively recommendation made by the investiga- Undoubtedly, implementation of safety be confirmed to be completely extin- tion team, the deadline for compliance improvement measures often results in guished, the safest course of action 24 • July-September 2019 ISASI Forum typically requires the earliest possible and engine area. MFOHE. The fuel disagree checklist descent, landing, and evacuation. The suggested four scenarios in which a fuel • Cabin crew manual also recommends that pilots leak should be suspected and thus the The cabin crewmembers could have had should utilize all available sources fuel leak checklist should be performed. a view of the fire situation at the right of information in making a decision One such scenario is when the totalizer wing and engine area through the cabin regarding evacuation. The manual fuel quantity is less than the calculated windows. The flight crew could have asked also highlights that key factors to be fuel quantity. Given that the totalizer the cabin crewmembers what they could considered include the urgency of the fuel quantity was about 79 tonnes and see. According to the PIC, the flight crew situation (e.g., possibility of signifi- the calculated fuel quantity about 83 was aware that cabin crew members were cant injury or loss of life if a signifi- tonnes, the flight crew should have cant delay occurs). The manual also a source of information throughout the occurrence. However, the flight crew was concluded that it had to proceed on to recommends that, in case of doubt, the fuel leak checklist. an evacuation should be considered. not able to attend to every call from the cabin crew as it had to prioritize tasks. In As mentioned, the flight crew’s own • On the other hand, the operator’s terms of obtaining information on the fire, assessment and fuel calculation put flight crew training manual also the flight crew gave priority to the task the remaining fuel quantity at about recognizes that fire may be spread- of communicating with the FC as he was 79 tonnes, which tallied well with the ing rapidly from spilled fuel or other the subject-matter expert and would have totalizer fuel quantity figure. This gave flammable materials, which may a better assessment of the fire from his the flightcrew members the confidence endanger the people who have left location outside the aircraft. that the totalizer fuel quantity was ac- the aircraft or are still on the escape Admittedly, the situation that the flight curate and they were not experiencing slides. crew faced was a stressful one. In trying a fuel leak. They decided that the fuel The flight crew will have to balance the to decide whether to evacuate, the crew disagree message was a spurious one pros and cons of a decision to evacuate was further disadvantaged as there was and that, therefore, there was no need given the situational picture that it has. It no indication of fire from the aircraft’s to conduct a fuel leak check. cannot be overemphasized that the flight fire detection system. The fire detection During the initial training to oper- crew needs to exhaust all possibilities and elements were shielded from the fire by ate this aircraft, the operator provides all available resources to try to build a the engine cowlings, and there were no training to all its pilots to understand situational picture that is as accurate as fire detection elements located outside the requirements of the fuel disagree possible. the engine cowling. There was also no checklist. However, in this case, the In this occurrence, there were a number guidance in the manuals available to the flight crew appeared to have misin- of resources that were not used by the pilots when there is a reported fire, but terpreted certain requirements of this flight crew but that could have been of the aircraft fire detection systems did not checklist even though it has undergone help. indicate so. the training. In that situation, the flightcrew mem- • Taxiing camera system As part of its recurrent training for bers depended on the FC as the sole The aircraft was equipped with a system pilots, operators may wish to consider source for information collection, and it of cameras installed at various locations including periodic refresher training on may have slipped their minds to consider to assist the flight crew in taxiing. If the requirements of the checklists that switched on, the system could have pro- alternative ways of gathering information. are used infrequently. This will increase Research has shown that decision-making vided real-time video of the exterior of the the likelihood that checklists will be under stress may become less systematic fuselage. There was one camera installed executed as intended. on the leading edge of the right horizon- and more hurried and that fewer alter- tal stabilizer. This camera could provide native choices are considered. The flight the flight crew with a vantage view of the crew’s behavior was consistent with the Conclusion fire. According to the flight crew, it would research findings. This occurrence brought several inter- usually switch on this camera system It is recognized that it may not be possi- esting issues into consideration. Other when taxiing the aircraft, as required by ble for an operator to practice its pilots on than having to determine the cause and the operating procedures. However, in this checklist response for all possible emer- propagation of the fire, other factors occurrence, the crew did not switch on the gency and abnormal situation scenarios. It such as regulatory requirements, risk system because it had not reached the tax- is therefore all the more critical that pilots management processes, recurrent develop the ability to always consider al- iing phase as the crew had been instructed training for flight crews, and human ternatives and other resources when they by ATC to stop at the intersection between performance were considered in this encounter a situation that is not dealt the runway and rapid-exit Taxiway E7. investigation. with by any checklist. • Cockpit escape window In conclusion, the experience and The flight crew could have opened the lessons learned through this investiga- cockpit escape window on the right side Execution of infrequently tion will benefit stakeholders in the to find out the situation outside. Extend- used checklist aviation industry as they continue to ing the upper body out of the right escape The fuel disagree message that the flight seek the optimal balance between window would allow a person to obtain a crew encountered was a result of the fuel ensuring safety, maintaining operating view of the fire situation at the right wing leak after the tube had cracked in the efficient, and profitability. July-September 2019 ISASI Forum • 25 ISASI Kapustin Scholarship Essay HOW AN EXTREME SPORT The following article is the final of four essays from the 2018 HIGHLIGHTS BROADER Kapustin scholarship winners. The number of scholars select- ISSUES FOR AIR SAFETY ed each year depends upon the amount of money ISASI mem- INVESTIGATORS: bers donate annually to the scholarship fund. Details about SKYDIVING OPERATIONS AND scholarship applications and additional information can be AIR SAFETY INVESTIGATION found on the ISASI website at By Avery Katz, Embry-Riddle Aeronautical University www.isasi.org. Application and essay deadlines are mid-April of t is no secret that in the United The following evidence will show each year—Editor. States, National Transportation a major disconnect between those Safety Board (NTSB) investigators who investigate accidents and those Ioften have their hands full. According who make regulations based upon the to the Federal Aviation Administra- findings. Thus, a significant challenge to tion (FAA), there were 513 accidents air safety investigators is a lack of useful between the years of 2013 and 2016 that regulations originating from accident were investigated by the NTSB, which investigations. has about 400 employees (Fact Sheet 2014). Despite this workload, the NTSB Accident review (in chronological still does a fantastic job. order) As a licensed skydiver with more than The first accident of the four, begin- 300 jumps, a former Cessna jump pilot, ning in the 2004–2016 period studied, and president of the Aviation Safety Ad- occurred on Oct. 24, 2004. The aircraft visory Council at Embry-Riddle, I have involved was a Cessna 206, tail num- always paid a great deal of attention to ber N8619Z (Rep. No. CHI05LA014). jump plane accidents and the subse- Generally, skydiving aircraft will climb quent NTSB reports. A website called to various altitudes depending upon Diver Driver has done a fantastic job of aircraft type. For instance, PAC 750s Avery Katz chronicling all aviation accidents since generally climb to 13,500, Cessnas nor- 2018 Kapustin scholarship recipient 1982. mally climb to 10,500, etc. In this case, Last year when reviewing these the 206 climbed to the normal altitude reports on the site, I noticed a trend. of 10,500 feet. Since 2004, there have been approx- The NTSB stated, “Aircraft control imately 21 fatal accidents directly [was] not possible by the pilot fol- related to jump plane operations (Jump lowing a premature deployment of a Plane, 2018). What is alarming? Nearly parachute as a parachutist exited the 20 percent of the accidents share a jump airplane during cruise flight. The common cause that has resulted in five inverted spin encountered by the pilot fatalities. This is the result of, as the was an additional cause.” (Rep. No. NTSB states, “inadvertent deployment CHI05LA014) of the reserve parachute.” (Rep. No. In this case, we see a simple inadvert- ERA09LA435) ent deployment by the jumper. With A review of the evidence will show no further information listed, the likely that despite a trend in probable cause cause was that the reserve handle was across four accidents, there has been no jumped on exit either by the jumper or Avery Katz presents his 2018 essay to the change in verbiage within the FAR AIM by contact with the airframe. participants of ISASI 2018 or additional flight training require- The second accident occurred on ments to help prevent future incidents. April 19, 2008. The aircraft was again Unfortunately, this means similar a Cessna 206, tail number N2537X accidents will likely occur in the future, (Rep. No. DEN08FA078). This was an thus increasing the workload of investi- interesting case in which multiple gators who have already presented the problems ultimately led to the fatal evidence needed to decrease accident accident. While at altitude, the pilot figures. entered a stall and spin when trying to

26 • July-September 2019 ISASI Forum position for the jump run. This led to a seatbelt to the handle that released the certainly provides clear guidance on rapid, uncontrolled descent during which reserve parachute. Therefore, the reserve reserve parachutes and how often they multiple jumpers were able to bail out of parachute deployed when the passenger need to be repacked. It even goes into the airplane. moved.” (Rep. No. CEN13LA500) detail to discuss varied materials, canopy However, sometime during the de- The NTSB noted the probable cause conditions, etc. Yet in the entirety of the scent, witnesses claimed it appeared the as “The improper routing of the seat- FAR AIM, there is no mention of specific airplane leveled out at “1,000 to 5,000 belt, which resulted in the inadvertent requirements for the construction of a feet,” indicating that the pilot may have deployment of the reserve parachute, and reserve deployment handle, even though regained control. However, a passenger’s the open jump door, which allowed the there is clearly a trend that certain de- reserve parachute could then be seen passenger to be pulled from the airplane.” signs lead to fatalities on board aircraft. “wrapped around the tail,” indicating an (Rep. No. CEN13LA500) This probable Furthermore, given the high percentage inadvertent deployment. cause is very telling. While it does not of Cessna 206s in accidents such as those At this point, the aircraft again became specifically list D-handles, the process examined, one must ask if the FAA has uncontrolled and “spun or dove to the of elimination can be used to figure out implemented any kind of special training ground.” (Rep. No. DEN08FA078) The which type of handle was involved. Gen- for jump pilots in the aircraft. The answer NTSB report stated, “Contributing factors erally, there are D-handles and spongy, is no. It has been proven that special fed- in this accident were the entanglement of soft fabric handles. However, D-handles eral regulations (SFAR) have been effec- the parachute in the elevator control sys- are the only type with a hold point, thus tive in the past. For instance, it reduced tem, reducing the pilots ability to regain they are the only type that could have had MU-2 accidents from 40 fatalities between control.” (Rep. No. DEN08FA078) While a seatbelt threaded through them. 1988 and 2008 to only two between 2008 the reserve deployment was not the first and 2016 (new MU-2B, 2016). Despite this, event in the sequence that led to this The trend the FAA has not implemented any such accident, it certainly didn’t make recovery Between the years of 2004 and 2016, the training for Cessna 206s converted for any easier for the pilot. In fact, witness NTSB positively identified inadvertent jump operations. testimony suggests the pilot may have reserve deployments as the cause of 19 recovered but then entered a secondary percent of fatal aircraft accidents in the Conclusion and recommendation spin upon reserve deployment. skydiving community. Upon closer exam- The findings in recent aircraft accidents The third accident occurred on Aug. ination, nearly 10 percent were directly show a trend, yet the FAA has failed to 1, 2009. The aircraft type was a Beech- related to the use of a D-handle. Curious- input new regulations in response to the craft B90, tail number N1999G (Rep. ly, a further statistic is that 75 percent of trends that have been displayed over a No. ERA09LA435). The NTSB stated, the fatal accidents involving inadvertent 12-year period of jump aircraft operation. “An instructor positioned himself at deployments were Cessna 206 crashes. Probable cause findings in four accidents the door opening with his jump stu- These are proven statistics gained from have identified the reserve parachute as dent nearby. The student inadvertently numerous NTSB reports. Thus, the ques- a factor, and at least two of them were pulled the instructor’s reserve parachute tion that must be answered is: Has the further tied to a specific D-ring design. D-ring, deploying the chute and pulling FAA made any changes to federal regula- Furthermore, 75 percent of these acci- the instructor out of the airplane. The tions that would affect reserve contain- dents involved the same kind of aircraft, a instructor contacted the left horizontal er/handle construction, ban the use of Cessna 206. stabilizer then descended toward the D-handles, or change training standards Yet this is where the work of an investi- ground coming to rest suspended in a tree for jump aircraft with higher fatality gator stops, and another organization by his parachute.” (Rep. No. ERA09LA435) rates? The answer is no. such as the FAA must take charge. The While blame can certainly be placed on organization must review the information the competency of the student in this Federal regulations provided and create meaningful change instance, it is also questionable that an The section of the FARAIM that deals in regulations to save lives. In the example instructor was using an easily catchable with reserve parachutes is 14 CFR 105.43, of skydiving operations used throughout reserve D-handle when jumping with an “Use of single-harness, dual-parachute this paper, meaningful change would inexperienced student. systems.” The following excerpt is taken include an addition to the verbiage in Part The final accident occurred on Aug. from that section: “The main parachute 105. Examples could include a ban of 16, 2013. The aircraft type was a Cessna must have been packed within 180 days D-rings, which potentially serve as a snag 206, and the registration was N2070K. before the date of its use by a certificat- hazard on exits and seatbelts and can The NTSB stated, “During the flight, [a] ed parachute rigger, the person making inadvertently be deployed by inexperi- passenger moved forward in the cabin, the next jump with that parachute, or a enced student jumpers. Furthermore, an which resulted in the passenger's reserve noncertificated person under the direct SFAR could be implemented that would parachute inadvertently deploying and supervision of a certificated parachute help future Cessna 206 jump pilots create the passenger being pulled through the rigger. (b)The reserve parachute must a safer environment for skydivers. The open jump door. The passenger hit the have been packed by a certificated para- most important fact is that this issue doorframe, and the parachute became chute rigger….” (CFR 14, § 105.43 (2018)) spans beyond skydiving. The aviation entangled with the empennage.… A post- In Part 105, this is the only mention of system in the United States could benefit accident examination revealed that the reserve parachutes in a sport container passenger had inadvertently attached his (excluding tandem rigs). The FAA (Continued on page 30)

July-September 2019 ISASI Forum • 27 NEWS ROUNDUP

ESASI Holds 2019 Seminar at Rolls-Royce Center icana de Servicios de Navegación Aérea); and Dr. Leonidas Duarte (Instituto Nicaragüense de Aviación Civil). The European Society of Air Safety Investigators (ESASI) work- shop was run on three sessions: • How can the investigation of design aspects be enhanced to improve safety throughout the lifecycle of an aircraft? • How effective are safety recommendations and safety ac- tions related to aircraft design? • How can we further improve the relationship between investigators from safety investigation authorities, manufac- turers, regulators and operators? At the start of each session, there was an introductory state- ment followed by a number of presentations representing views from across the industry. Table discussions followed, and each session was concluded with an open discussion session. The main points were captured by the 14 table rapporteurs. Seating was preassigned to ensure that each table had a wide representation from across the aerospace industry. The ESASI committee is reviewing the views of the workshop A delegation of Nicaraguan aviation air safety and corporate officials meets in Buenos Aires, Argentina, with ISASI's Latin American Regional that will be used to produce a briefing paper that will be widely Society representatives. distributed. The 14 rapporteurs will be invited to review the draft paper.

SASI Pakistan Holds Reachout Sessions

The Pakistan Society of Air Safety Investigators hosted a series of courses, workshops, and presentations as the 54th ISASI Reachout in Karachi, Lahore, Kamra, and Islamabad, Pakistan, on June 10–25. The arrangements and coordination were made by Wing Commander (Ret.) Syed Naseem Ahmed, president of SASI Pakistan. Pakistan International Airlines (PIA) supported the courses by providing air transport for the two instructors from Toronto, Canada, to Karachi and return from Lahore. Capt. Mohsin Ausaf Khan, the PIA senior manager for safety, extended his full support, without which it wouldn’t have been possible to conduct these Reachout events. The one-week course (incident investigation and safety risk management) was conducted for PIA with some participants from the Pak Army Aviation, the Pak Navy, and the Pakistan Air Participants at the European Society of Air Safety Investigators workshop take time out for a photo at the Rolls-Royce Learning and Development Force at the PIA Training Centre in Karachi from June 10–14. Centre. The ISASI instructors were Mike Doiron, representing ISASI and Cirrus Aviation (Halifax, Nova Scotia, Canada), and ISASI Inter- national Councilor Caj Frostell. The Pakistan Society coordination and arrangements were Latin America Regional Society Meets with Nicaraguan handled by Syed Naseem Ahmed and Mohsin Ausaf Khan. The Delegation SASI Pakistan officials took great care to ensure that the instruc- tors and participants had everything they needed to maximize learning and course success. Instructors prepared master copies Air safety and corporate officials from Nicaragua meet recently with officials from the Latin American Society of Air Safety of their training material, and the Pakistan Society arranged Investigators (LARSASI) in Buenos Aires, Argentina, to discuss for reproduction of the presentations in the form of a hardcopy regional air safety issues and to join the ISASI regional society. participant handout. Shown from the left are Ing. Guillermo Guido, director, ANIA PIA Flight Safety Seminar—On the afternoon of June 14, (Asociación Nicaragüense de Investigación de Accidentes); instructors and course participants were invited to attend the Enriqueta Zambonini, LARSASI advisor; Capt. Carlos Salazar, PIA Flight Safety Seminar at the PIA Training Centre auditorium. DGAC Nicaragua/Pte. Griaa (Grupo Regional de Investigación The seminar was organized by Mohsin Ausaf Khan and included Accidentes de Aviación Civil); Capt. Daniel Barafani, Pte. LARSASI; several presentations on safety issues by PIA Safety Department Dra. Eveling Arauz, advisor, COCESNA (Corporación Centroamer- experts, as well as presentations by ISASI instructors. 28 • July-September 2019 ISASI Forum NEWS ROUNDUP

Reachout participants at ATS Karachi on June 20. CAA Additional Director Safety Management and Quality Assurance Hasan Mujahid is in the center with ISASI instructors.

Karachi Institute of Economics & Technology—On June 17, An evening social event was hosted by SASI Pakistan for its mem- instructors gave half-day presentations to the aviation program bers and instructors at the Monal restaurant on a mountainside students of the Pakistan Air Force Karachi Institute of Econom- overlooking Islamabad. ics & Technology. The event was held at the PIA Training Centre On June 2, SASI Pakistan and the Pakistan Air Force invited auditorium. instructors for a tour four hours northeast of Islamabad into the CAA Airworthiness Directorate—On June 18–19, instructors foothills of the Himalaya and visited a Pakistan Air Force base in were invited to the civil aviation authority (CAA) Pakistan Kalabagh, close to the border with India. Airworthiness Directorate to conduct a two-day workshop for Numerous seminar participants noted that it was a unique invited operator representatives and airworthiness inspectors. opportunity arranged by SASI Pakistan in which the aviation Air Traffic Services in Karachi—On June 20, the CAA Air Traffic industry (CAA, airlines, Air Force) came together with road safety Services Department invited instructors to conduct a workshop authorities and road safety experts to discuss safety issues. From for senior air traffic controllers in Karachi. The event was held an ISASI instructor perspective, the multitude of ISASI Reachout at the CAA ATS premises. In the evening, instructors flew from activities was a unique opportunity to exchange experiences and Karachi to Lahore. discuss different ways of implementing safety strategies, han- Air Traffic Services in Lahore—On June 21, the CAA Air Traffic dling emergency situations, carrying out investigations, imple- Services in Lahore invited instructors to conduct a workshop menting safety actions, and exchanging ideas. Instructors truly for senior air traffic controllers. The event was held at the CAA appreciated the excellent arrangements, the interactions with premises. Asad Gondal, the CAA’s senior air traffic controller, was management and course/workshop participants, and the ISASI’s contact and support person in Lahore. He arranged an exceptional hospitality. interesting visitor tour of Lahore for the instructors. On June 22, instructors took a bus from Lahore to Islamabad. Pakistan Aeronautical Complex Kamra—On June 24, instruc- In Memoriam tors presented a half-day workshop at the Pakistan Aeronautical Robert Gould died at home in Florence, Massachusetts, Complex (PAC) in Kamra. The PAC Kamra is an impressive aer- USA, on June 4 of cardiac arrest. Robert joined ISASI in 2013. onautical manufacturing facility that produced mostly military The Daily Hampshire Gazette noted the he was involved with aircraft, glass cockpits, and aircraft instrumentation and has aviation from an early age and traveled with his parents in the family plane to Alaska, throughout the western United States, some 30,000 employees. The PAC Kamra hosts were Air Commo- and to the Bahamas. Robert worked at an aircraft engine facil- dores Waqas Mahni and Asif Maqsood. ity and had a pilot license before he was old enough to obtain a SASI Pakistan International Safety Awareness Seminar—In driver’s license. preparation for the SASI Pakistan International Safety Awareness After graduating from Kent State University and earning a Seminar titled “Future of Accident Investigations” on June 24, in- master’s degree at SUNY Empire State College in New York, structors participated in a press conference to promote the next Robert joined the U.S. Navy and served active duty at the Patux- day’s seminar. The seminar was attended by 400 participants, ent River Naval Air Station in Maryland. He then joined the SASI Pakistan members, safety experts representing aviation and Naval Air Systems Command as an active reservist where he road safety, and students from the Air University in Islamabad. retired at the rank of captain in 2002. He worked for business aviation maintenance companies for many years. The seminar was held in the auditorium of the Air University. Beginning a teaching career, Robert retired to form his own The inauguration was delivered by the Vice Chief of Air Staff Air consulting company. He taught aviation safety courses around Marshal Asim Zaheer. The program included presentations from the world and was an instructor for the University of Southern the two ISASI instructors and a number aviation and road safety California Aviation and Security Program. Robert was an active specialists, including Squadron Leaders Mehmood Masood and volunteer for Wright Flight, an organization that introduces Faisal Bashir Bhura. young people to aviation fields, and served on the boards of The seminar was a major success for SASI Pakistan, Syed Na- several aviation organizations. He received several awards seem Ahmed, and SASI Pakistan members Mohsin Ausaf Khan, during his careers, including the Charles Taylor Award from the Major General (Ret.) Muhammad Azam (vice president of SASI U.S. Federal Aviation Administration for lifetime achievement Pakistan), Squadron Leader Athar Sajid, and Dr. Fatima Yousuf. in aviation maintenance. July-September 2019 ISASI Forum • 29 ISASI Kapustin ISASI INFORMATION Scholarship Essay OFFICERS (Continued from page 27) President, Frank Del Gandio ([email protected]) in many instances from seeing more changes based upon the NTSB’s findings. Executive Advisor, Richard Stone In this case, skydiving is simply indicative of a much larger issue. If this ([email protected]) Vice President, Ron Schleede problem can be solved in jump plane operations, it would create a pattern of ([email protected]) meaningful change throughout the aviation community. Secretary, Chad Balentine ([email protected]) Treasurer, Robert MacIntosh, Jr. References ([email protected]) COUNCILORS CFR 14, § 105.43 (2018). Australian, Richard Sellers ([email protected]) Fact Sheet–General Aviation Safety. (Sept.19, 2014). Retrieved Canadian, Barbara Dunn ([email protected]) April 9, 2018, from https://www.faa.gov/news/fact_sheets/news_ European, Rob Carter story.cfm?newsId=21274. ([email protected]) International, Caj Frostell ([email protected]) Jump Plane Accidents by Year (U.S.). (n.d.). Retrieved April 6, 2018, New Zealand, Alister Buckingham from http://diverdriver.com/accidents-by-year. ([email protected]) Pakistan, Wg. Cdr. (Ret.) Naseem Syed New MU-2B training rules published. (Sept. 7, 2016). Retrieved Ahmed ([email protected]) April 9, 2018, from https://www.aopa.org/news-and-media/all- United States, Toby Carroll news/2016/september/07/new-mu-2b-training-rules-published. ([email protected])

https://www.aopa.org/news-and-media/all-news/2016/septem- NATIONAL AND REGIONAL ber/07/new-mu-2b-training-rules-published. SOCIETY PRESIDENTS AsiaSASI, Chan Wing Keong (Rep. No. CEN13LA500). (n.d.). Retrieved April 10, 2018, from ([email protected]) https://www.ntsb.gov/_layouts/ntsb.aviation/brief.aspx?ev_ Australian, Richard Sellers id=20130822X63331&key=1. ([email protected]) Canadian, Barbara Dunn ([email protected]) European, Olivier Ferrante (Rep. No. CHI05LA014). (n.d.). Retrieved April 10, 2018, from ([email protected]) https://www.ntsb.gov/_layouts/ntsb.aviation/brief.aspx?ev_ Korean, Dr. Tachwan Cho (contact: Dr. Jenny id=20041028X01721&key=1. Yoo—[email protected]) Latin American, Daniel Barafani, PTE ([email protected]) (Rep. No. DEN08FA078). (n.d.). Retrieved April 10, 2018, from Middle East North Africa, Khalid Al Raisi https://www.ntsb.gov/_layouts/ntsb.aviation/brief.aspx?ev_ ([email protected]) id=20080429X00561&key=1. New Zealand, Paul Breuilly ([email protected]) (Rep. No. ERA09LA435). (n.d.). Retrieved April 10, 2018, from Pakistan, Wg. Cdr. (Ret.) Naseem Syed https://www.ntsb.gov/_layouts/ntsb.aviation/brief.aspx?ev_ Ahmed ([email protected]) id=20090803X05217&key=1. Russian, Vsvolod E. Overharov ([email protected]) United States, Toby Carroll ([email protected]) UNITED STATES REGIONAL CHAPTER PRESIDENTS Alaska, Craig Bledsoe ([email protected]) MOVING? NEW E-MAIL ACCOUNT? Arizona, Bill Waldock ([email protected]) Dallas-Ft. Worth, Erin Carroll Do you have a new mailing address? Have you recently changed ([email protected]) Great Lakes, Matthew Kenner your e-mail address? Then contact ISASI at [email protected] to ([email protected]) ensure that your magazine and other ISASI materials are delivered Mid-Atlantic, Frank Hilldrup ([email protected]) to you. Please include your previous address with your change Northeast, Steve Demko ([email protected]) request. Members in Canada, New Zealand, and Australia should Northern California, Kevin Darcy ([email protected]) contact your national society. Pacific Northwest, (Acting) John Purvis ([email protected]) Rocky Mountain, David Harper ([email protected])

30 • July-September 2019 ISASI Forum ISASI INFORMATION

Southeastern, Robert Rendzio Air Astana JSC Global Aerospace, Inc. ([email protected]) Air Canada Grup Air Med S.A. Southern California, Thomas Anthony Air Canada Pilots Association Grupo Regional de Investigación de Accidentes de Air Line Pilots Association Aviación ([email protected]) Airbus Gulfstream Aerospace Corporation Airclaims Limited Hall & Associates LLC COMMITTEE CHAIRMEN Air New Zealand Hawaiian Airlines Airways New Zealand HNZ New Zealand Limited Audit, Dr. Michael K. Hynes All Nippon Airways Co., Ltd. (ANA) Hogreen Air ([email protected]) Allianz Honeywell Aerospace Award, Gale E. Braden ([email protected]) Allied Pilots Association Hong Kong Airline Pilots Association Ballot Certification, Tom McCarthy Aloft Aviation Consulting Human Factors Training Solutions Pty. Ltd ([email protected]) Aramco Associated Company Independent Pilots Association Insitu, Inc. Board of Fellows, Curt Lewis ([email protected]) Asiana Airlines Asociación Nicaragüense de Investigación de Interstate Aviation Committee Bylaws, Darren T. Gaines Accidentes Irish Air Corps ([email protected]) ASPA de Mexico Irish Aviation Authority Code of Ethics, Jeff Edwards ([email protected]) ASSET Aviation International Pty. Ltd. Japan Transport Safety Board Membership, Tom McCarthy ([email protected]) Association of Professional Flight Attendants Jones Day Mentoring Program, Anthony Brickhouse Australian and International Pilots’ Association KLM Royal Dutch Airlines Korean Air ([email protected]) (AIPA) Australian Transport Safety Bureau Korea Aviation & Railway Accident Nominating, Troy Jackson Aviation Investigation Bureau, Jeddah, Investigation Board ([email protected]) ­ Kingdom of Saudi Arabia L-3 Aviation Recorders Reachout, Glenn Jones ([email protected]) Aviation Safety Council Learjet/Bombardier Aerospace Scholarship Committee, Chad Balentine Avisure Lion Mentari Airlines, PT ([email protected]) Azure Aero Ltd Lockheed Martin Aeronautics Company Middle East Airlines Seminar, Barbara Dunn ([email protected]) Becker Helicopters Pty. Ltd. Bell Midwest University Bundesstelle fur Flugunfalluntersuchung (BFU) Military Air Accident Investigation Branch WORKING GROUP CHAIRMEN Bureau d’Enquêtes et d’Analyses (BEA) Military Aircraft Accident & Incident CAE Flightscape Investigation Board Air Traffic Services, Scott Dunham (Chair) Ministry of Transport, Transport Safety ([email protected]) Cathay Pacific Airways Limited Centurion Aerospace Ltd. Investigation Bureau, Singapore Ladislav Mika (Co-Chair) ([email protected]) Charles Taylor Aviation National Aerospace Laboratory, NLR Airports, David Gleave ([email protected]) China Airlines National Institute of Aviation Safety and Cabin Safety, Joann E. Matley Civil Aviation Authority, Macao, China Services ([email protected]) Civil Aviation Department Headquarters National Transportation Safety Board Civil Aviation Safety Authority Australia National Transportation Safety Committee- Corporate Affairs, Erin Carroll Indonesia (KNKT) ([email protected]) Civil Aviation Safety Investigation and Analysis Center NAV CANADA Critical Incident Stress Management (CISM), Colegio Oficial de Pilotos de la Aviación Ocean Infinity David Rye--([email protected]) Comercial (COPAC) Pakistan Air Force-Institute of Air Safety Flight Recorder, Michael R. Poole Pakistan Airline Pilots’ Association (PALPA) Commercial Aircraft Corporation of China Pakistan International Airlines Corporation (PIA) ([email protected]) Cranfield Safety & Accident Investigation Papua New Guinea Accident Investigation General Aviation, Steve Sparks Centre Commission (PNG AIC) ([email protected]) Curt Lewis & Associates, LLC Parker Aerospace Co-Chair, Doug Cavannah Dassault Aviation Petroleum Air Services DDAAFS Phoenix International Inc. ([email protected]) Defence Science and Technology Organisation Government Air Safety Facilitator, Plane Sciences, Inc., Ottawa, Canada (DSTO) Pratt & Whitney Marcus Costa ([email protected]) Defense Conseil International (DCI/IFSA) PT Merpati Nusantara Airlines Human Factors, Edma Naddof Delft University of Technology Qatar Airways ([email protected]) Delta Air Lines, Inc. Republic of Singapore Air Force (RSAF) Investigators Training & Education, Directorate of Flight Safety (Canadian Forces) Rolls-Royce PLC Discovery Aur Defence Royal Danish Air Force, Tactical Air Command Graham R. Braithwaite Dombroff Gilmore Jaques & French P.C. ([email protected]) Royal Air Force DRS C3 & Aviation Company, Avionics Line of Royal New Zealand Air Force Military Air Safety Investigator, James W. Roberts Business RTI Group, LLC ([email protected]) Dubai Air Wing Saudia Airlines-Safety Unmanned Aerial Systems, Tom Farrier Dubai Civil Aviation Authority Scandinavian Airlines System ([email protected]) Dutch Airline Pilots Association Sikorsky Aircraft Corporation Dutch Safety Board Singapore Airlines Limited Eclipse Group, Inc. Southern California Safety Institute CORPORATE MEMBERS Education and Training Center for Aviation Southwest Airlines Company AAIU, Ministry of Transport Safety Southwest Airlines Pilots’ Association Abakan Air EL AL Israel Airlines Spanish Airline Pilots’ Association (SEPLA) Accident Investigation Board (AIB) Army Aviation Embraer-Empresa Brasileira de Aeronautica State of Israel Accident Investigation Board Norway S.A. Statens haverikommission Accident Investigation Bureau Nigeria Embry-Riddle Aeronautical University Swiss Accident Investigation Board (SAIB) Administration des Enquêtes Techniques Etihad Airways The Air Group Adnan Zuhairy Engineering Consultancy European Aviation Safety Agency (EASA) The Boeing Company Aegean Airlines EVA Airways Corporation The Japanese Aviation Insurance Pool (JAIP) Aer Lingus Executive Development & Management Advisor Transportation Safety Board of Canada Aero Republica Finnair Plc Turbomeca Aerovias De Mexico, S.A. De C.V. Finnish Military Aviation Authority Ukrainian National Bureau of Air Accidents and Agenzia Nazionale Per La Sicurezza Del Volo Flight Data Services Ltd. Incidents of Civil Aircraft Air Accident Investigation Authority of Hong Kong Flight Data Systems Pty. Ltd. UND Aerospace Air Accident Investigation Bureau of Mongolia Flight Safety Foundation United Airlines Air Accident Investigation Bureau of Singapore Fugro Survey Middle East Ltd. United States Aircraft Insurance Group Accident Investigation Committee of Thailand Gangseo-gu, Republic of Korea University of Balamand/Balamand Institute of Air Accident Investigation Unit-Ireland GE Aviation Aeronautics Air Accident Investigation Sector, GCAA, UAE General Aviation Manufacturers Association University of Southern California Air Accidents Investigation Branch-UK German Military Aviation Authority, Directorate of Virgin Galactic Air Asia Group ­ Aviation Safety Federal Armed Forces WestJet July-September 2019 ISASI Forum • 31 ISASI 107 E. Holly Ave., Suite 11 Sterling, VA 20164-5405 USA

INCORPORATED AUGUST 31, 1964 CHANGE SERVICE REQUESTED

ANZSASI Holds Annual Regional Seminar

ew Zealand Councilor Alister Buckingham reports that the 2019 NAustralian/New Zealand Society of Air Safety Investigators (ANZSASI) regional seminar was held in Wellington, New Zealand, June 7–9. Some 93 delegates Attendees of the Asia-Pacific Cabin Safety Working Group meeting, which was hosted by the participated. The event teed New Zealand civil aviation authority. off on Friday with the inaugu- ral Trans-Tasman golf chal- on the hazards associated through the weekend with a business meeting was recon- lenge, with both sides fielding with lithium batteries. total of 20 first-class technical vened briefly at the end of some big hitters and a mixed To round off the day’s ac- papers plus a summary of the the Sunday morning session bag of others with varying tivity, 97 guests attended the cabin safety proceedings. Most to confirm (very gratefully) degrees of skill. The Kiwis ANZSASI welcome reception papers are now available on the Eastwood’s appointment. carried the day and secured at the conference hotel, where ASASI website. Other agenda items con- the trophy. there was no shortage of food Both societies conducted firmed that the Society was Also on that Friday the New and drink. This event was their business meetings im- in good physical and finan- Zealand civil aviation author- hosted entirely by NZSASI, mediately after the Saturday cial health, with the mem- ity (CAA) hosted an Asia-Pa- with the cost having been fac- program. The NZSASI agenda bership steady around the cific Cabin Safety Working tored into the seminar budget. included the confirmation 70 mark, and that there were Group meeting, and a record The reception had its own of the nominations for the sufficient funds in hand to 67 delegates attended. Pres- highlight—no speeches! Society’s Executive. Paul Breu- cover a total loss on a region- entations ranged from lessons Despite the hectic start to illy was confirmed as the new al seminar if that extreme learned from incidents, the weekend, the seminar was president, Louise (Lou) Child circumstance were ever to reporting culture, dangerous anything but an anticlimax. as vice president, and Alister arise. This year’s seminar goods, and aviation security First up was Björn Hennig, Buckingham reconfirmed as was again budgeted to incur to sexual assault on board air- who traveled from Germany councillor. These were the only a small loss, in keeping with craft and the airline response. especially for the seminar, his nominations received, as there the Society’s not-for-profit Topical issues were also second such appearance. Hen- was none for secretary/treas- status. discussed, including hand lug- nig delivered the Ron Chippin- urer up to this point. An appeal The ritual president- gage, fatigue, and passenger dale Memorial Presentation; was made to the membership to-president handover of the comfort devices. A highlight his theme was data-driven for someone to step up and seminar cowbell to ASASI of the meeting was a presenta- fatigue management. Bur- relieve Russell Kennedy after was the final act of the tion by Michael Burdick, a U.S. dick made a reprise appear- his 20-odd years of sterling ser- weekend, after Rick Sellers Federal Aviation Administra- ance, this time with a safety vice in the role. Early the next had announced the venue tion dangerous goods special- management system theme, morning, Michael Eastwood and dates for the 2020 ist, who provided an update and the seminar continued volunteered his services. The events. 32 • July-September 2019 ISASI Forum