The third part of a three-part story by Charlotte Adams
In a notice of proposed rulemaking (NPRM) last fall, the Federal Aviation Administration (FAA) underscored the importance of helicopter flight data monitoring (HFDM). The NPRM asked whether a lightweight aircraft recording system (LARS) mandate for air ambulance helicopters would increase the industry's participation in the agency's FDM program, known as flight operations quality assurance (FOQA). The FAA also is considering requiring air ambulance operators to adopt HFDM programs even though such a step would rule out protected data sharing.
The NPRM, observers say, reveals contradictions in the FAA's initial proposals about LARS. To come up with the best solution, choices need to be made that are not made there. Which is more important, for example, accident investigation or accident prevention? Low cost or sophisticated recording parameters? Simplicity or complexity in design and installation?
The NPRM raises other questions. Is video adequate to record flight control position inputs? And why the obsession with weight? The difference in weight that the FAA cites between a combi voice/data recorder and a LARS is fairly slight. Obviously the document is just the opening salvo in a longer engagement.
Editor's Note: the first article in this series was titled "FAA tests waters for helicopter emergency medical services recording with mixed response" and the second was titled "Helicopter FDM: affordable hardware"
FAA officials have already approved FOQA programs using lightweight, self-contained, limited-parameter devices. Could these programs be cancelled if LARS requirements escalate in a final rule? Would FOQA applicants after the final rule risk having to install a relatively more complex and expensive box than those already approved?
It is also interesting, according to one pilot and analyst, that a potential LARS requirement should be limited to air ambulance helicopters when, based on National Transportation Safety Board (NTSB) and Government Accountability Office (GAO) reports cited by the FAA, medical helicopters and airplanes have experienced a roughly equivalent accident rate. On another level, if analysis is as important as FAA says it is, how much analysis is enough?
Flight data analysis ranges from do-it-yourself software, which supplies tools for operators to use on their own, to hands-on services that provide customized reports and intervention advice. While flight data analysis is mature in the airlines, it is relatively new for helicopters, and there are just a few specialists. As HFDM becomes more common, however, this sector is sure to grow. If they can afford the technology, operators would gain insight into their flight operations and improve safety.
On the do-it-yourself side is a product expected to be released this year from General Aviation Safety Network (GASN) in St.-Hubert, Quebec, which introduced its Wi-Flight package for fixed-wing aircraft this year in 2010. The business model for the Wi-Flight HFDM recorder -- a smartphone with an inertial sensor board -- resembles that of wireless carriers, says Pascal Gosselin, GASN chief executive officer.
The aircraft hardware is about $800, so the company expects to recoup its development costs through subscriptions to its flight playback and analysis services. Although there are other costs, such as a $500 accelerator for faster uploads and a $200 Wi-Fi outdoor access point, the overall hardware outlay would be attractive for smaller operators with legacy helicopters. The monthly per-tail fee for unlimited flight analysis and playback -- plus three years of data storage -- has not yet been established, but will be based on quantity discounts for fleet size. (Existing Wi-Flight fixed-wing pricing ranges from $29 to $59.)
At the heart of the Wi-Flight HFDM is its Web-based, "automated analysis." Wi-Flight gives users the tools to drill down into their own flight data. The analysis software also uses three terrain elevation databases, inputs from the HFDM's internal barometric altimeter and weather information from METAR weather reporting stations along the flight path to derive AGL (above ground level) data.
The company plans two interfaces. One is a visual playback of events in a flight -- starting just before the occurrence -- using 3D Google Earth imagery, streaming audio, and a Google Maps-based 2D moving map of the flight path. The other is a natural language-based search engine for browsing individual flights by tail number, time period or type of event. Flight data can also be graphed by parameter pairs, such as vertical speed vs. ground speed or altitude vs. G loading. Contextual flight playbacks can be viewed by clicking on the link contained in an email alert. Fleetwide graphing of aggregate data is not yet available but will be added, based on user feedback.
Unlike some other products, Wi-Flight's HFDM -- measuring about 4.75 by 2.5 by 0.75 inches and weighing 5 ounces -- will not use a removable data card, as these can easily be lost. Instead the glare shield-mounted unit will store flight data and audio in the phone's internal 8-gigabyte Micro SDHC memory and upload the information automatically to the Internet over Wi-Fi as soon as the plane lands. Downloads, however, only occur when the aircraft has been powered off, ground speed is zero, and vibration is no longer detected. Then data is uploaded over Wi-Fi at as fast as 54 megabits per second using the accelerator or at speeds of around 1 megabit per second over ADSL. If an operator's base lacks high-speed wired Internet, one can use a MiFi-type solution from a wireless carrier.
The basic device combines GPS, Wi-Fi, and dual, 3-axis accelerometers. Two audio channels pick up radio chatter and the area microphone. The sensor board provides a 3-axis magnetometer and a 3-axis MEMS rate gyro, as well as the baro altimeter. "We're going to have a full-featured AHRS [attitude heading reference system], with pitch, roll and heading, as well as barometric altitude," Gosselin says. He says that the product will record 100 hours of data and 100 hours of audio.
GASN plans to alert for events such as low-level buzzing, restricted airspace penetration, noise infractions, G exceedances, unstable approaches, and hard landings. It aims to detect things like premature turns after departure, high sink rate, main rotor unloading, and low rotor RPM, as well as determine main rotor RPM (via audio analysis) and provide rotor balance/vibration analysis. The company also plans to provide position reporting over the Global System for Mobile communication (GSM) in countries like Canada where this is allowed.
Gosselin is well aware of the "false positives" issue -- FDM-reported events that turn out to be non-events. Although GASN does not provide highly customized analysis, it keeps an eye out for false alerts. "We get copied on all alerts generated by the system," he says. "If we think it's a false positive due to an unforeseen condition that we can improve upon, we will tweak the code, reanalyze the flight and clear it.
"We've also had customers report non-detection of what they thought was a violation," he adds, alluding to users of the fixed-wing package. In one case low-level buzzing was not detected because a private airstrip was located within 3 nautical miles of the aircraft. "In cases like that we can delete the private airstrip in our database in order to enable low-level buzzing detection in that area."
A small group of companies offers analysis services to operators who lack either the time or the expertise to manage their FDM programs. It includes organizations such as CAPACG in Daytona Beach, Fla., Sagem Avionics in Grand Prairie, Texas, and newcomer Helinalysis Ltd., of Aberdeen, England, which specializes in helicopter operations.
Capt. Mike Pilgrim, a helicopter pilot and FDM analyst with CHC Helicopter, launched Helinalysis this year on a personal-time, non-conflict-of-interest basis. Pilgrim sees a need for more than just automated analysis. Considering that helicopters have hundreds of moving parts and generate large amounts of static electricity, "you're just asking for false alerts." False alerts can also be generated by equipment limitations, such as a radio altimeter in a popular model that reverts to zero at altitudes over 2,500 feet. If analysis is automated, how are events to be "validated," or proven to be genuine, he asks. Validation is key to an FDM program, he says, and "you absolutely need an experienced analyst for that." FDM analysis also needs to deal with complex data, he says. A Super Puma, for example, may have 250 parameters and an S-92, more than 600.
"Aggregating the data is what FDM is all about," Pilgrim adds. That is why video falls short of the mark in a full FDM program, he notes. It can tell you about an individual flight but does not lend itself to aggregation. Nevertheless, Pilgrim adds, Bell Helicopter is working on a cockpit information recorder (CIR) which is to have software to read the cockpit instruments and provide digital information. Bell says it will "embed software for image interpolation; analog to digital." But the level of accuracy the CIR will manage to achieve remains to be seen, Pilgrim says.
CAPACG -- CAP Aviation Consulting Group -- is a FDM analyst in the helicopter niche that offers monthly and quarterly subscriptions as well as customized reports. The company uses Appareo and Sagem analysis tools, as well as GA-FDM, a bundled FDM hardware and software suite developed with Alakai Technologies in Hopkinton, Mass. CAPACG helped oil-and-gas support giant Bristow and small air ambulance operator, Arkansas Children's Hospital (ACH) to launch their FAA-approved FOQA programs. Like its peers, GA-FDM embeds a reporting tool that provides aggregate trend information. But it also supplies normalized event rates, a better reporting metric than simple percentages, says Kipp Lau, vice president of FDM services for CAPACG. Not all FDM software is set up to do that, he says.
ACH, which uses Appareo hardware and tools, has outsourced analysis and reporting to CAPACG, which has provided actionable information after in-flight events. Early in the operator's experience with FDM, for example, the onboard equipment recorded an autopilot excursion during an IFR flight at night, recalls BJ Raysor, ACH's director of aviation. ACH was able to validate the occurrence as a yaw axis excursion somehow induced by aircraft computers. By analyzing the flight data, CAPACG was able to determine that no flight manual limitations had been exceeded, so that the aircraft was able to return to service without impacting the flight schedule of this two-ship operation. CAPACG also showed that it was an autopilot-induced event.
Startup Helinalysis, ironically, commenced services with one of the world's largest helicopter operators, the U.S. Army. Through a subcontract with Concurrent Technologies Corp. in Johnstown, Pa., CHC's Capt. Mike Pilgrim, Helinalysis' managing director and currently its only employee, reviewed the FDM program of the Army's Flight School XXI, Fort Rucker, Ala. Working evenings and weekends, using Appareo software, Pilgrim looked at individual and aggregate flight data from Army helos equipped with Appareo's GAU 2000 units.
Helinalysis "more than doubled" the number of triggers used by the Army flight school to identify events. That allowed service analysts to “get a much wider view of what's happening in their operation than they previously had," Pilgrim says. Helinalysis now is working with defense contractor, SAIC, on a 12-month project to advise the U.S. Army on its HFDM program.