Friday, September 1, 2006
Q&A: Lessons Learned From Las Vegas
Performance-based area navigation (RNAV) procedures allow more efficient and cost-effective use of the National Airspace System (NAS) by freeing aircraft from dependence on ground-based navaids. But these improvements don't come into being without considerable growing pains. The evolution of RNAV arrival and departure procedures at Las Vegas McCarran International Airport (LAS) is a good example. While LAS was not the first airport to attempt RNAV procedures in the terminal area, it was the first to be successful at it, although it took a two-year redesign effort to reach that point. But now that a firm foundation has been laid, the LAS procedures continue to evolve and the lessons learned have directly affected work at other airports. Avionics Magazine talked about the LAS experience with Brian Townsend, an Airbus A319 and A320 captain with the "new" US Airways and the NAS modernization chairman for the Air Line Pilots Association (ALPA). Townsend was deeply involved with the development, troubleshooting and implementation of the RNAV procedures at Las Vegas.
Avionics: Las Vegas is the only U.S. airport to offer RNAV departures and arrivals with the benefits of vertical profiles. How did the project get started and what was ALPA's role?
Townsend: The Four Corner Post plan was designed to address the air traffic and airspace inefficiencies resulting from increased demand at LAS. It's essentially a direction-based system that organizes aircraft from similar directions over a specific geographic position, or fix. Airspace is organized so that aircraft arriving from similar directions are directed over a specific fix and aircraft departing in similar directions are directed over a different fix. When LAS underwent their airspace redesign initiative in 2001, they had a clean slate to develop new arrival and departure procedures and routes. This was in the FAA's early stages of endorsing public RNAV procedures development for the National Airspace Redesign [NAR] effort.
I was the ALPA airport liaison representative for LAS and was approached by TRACON [terminal radar approach control] officials to help facilitate a meeting with key stakeholders for procedures development. I soon became involved and very interested in the RNAV procedures design concept, and provided the line pilot perspective for ALPA.
Avionics: There was a major learning curve at Las Vegas. Was McCarran the first airport to try this?
Townsend: In the late `90s to around 2000 there were a few airports that had developed RNAV STARs [standard terminal arrival routes] and SIDs [standard instrument departures]. They encountered problems similar to what we would later experience in Las Vegas, such as confusion with the charting and flying the procedures. It concerns training issues with pilots and controllers. The trend at the time tended to be, "if there's a problem, turn it off." It was unfortunate, but that's basically what occurred. When Las Vegas came out with their first product in 2001, and we encountered some procedure design and implementation issues, the initial reaction was to turn them off.
Fortunately, Las Vegas was going through a major redesign of their airspace in the Four Corner Post project. That worked to our advantage because the original procedures were no longer compatible with the redesigned airspace. So in some ways they were forced to keep the RNAV procedures. To the credit of the TRACON management at Las Vegas, they wanted to fix the problems and press on. They knew that RNAV was the wave of the future. So, as painful as it was, we held fast and forged ahead. Today, as a result, it works fine, so I guess that puts Las Vegas in the forefront. Even though they weren't the first airport, they were the first successful airport.
Avionics: RNAV is a lot more difficult in the terminal area than it is en route?
Townsend: RNAV is not new. People thought, since we'd been doing RNAV in the en route environment for 30 years, integrating RNAV into the terminal environment should be rather effortless. Pilots knew how to operate flight management computers. But it was more challenging in the terminal environment. There are more things going on closer to the airport, and there are altitude constraints and speed constraints that have to be met. There is also less time to react. So if a controller needed to amend a procedure--to change the altitude or airspeed restrictions, or to take a pilot off the procedure and then put him back on--that's where it became more of an operational challenge.
Avionics: What were the operational challenges that led to the redesign of the original procedures?
Townsend: The first phase of the arrivals into Las Vegas contained multiple runway transitions that essentially lead you to the approach end of the runway. One of the operational issues we had was with last-minute runway changes. That requires heads down in the cockpit, programming the computer. It's not a matter of a few seconds and a few key strokes to accomplish the task. It's a bit more labor intensive. So what was occurring was the pilots were essentially flying the wrong way. Either they didn't turn on the appropriate transition or they continued on a heading that was not part of the procedure. Initially the controllers attributed the errors to the procedures. It was also a training issue for the pilots.
Avionics: What did you do about it?
Townsend: A Pilot/Controller Procedures and Phraseology Work Group was formed to address and correct the issues.Now there are established procedures and guidelines for controllers and pilots to prevent this from happening. We determined that, to be safe, you need about 2 minutes to make the necessary programming changes for a runway transition amendment. It's also been made clear to the controllers that, if they need to make a last-minute change, they need to give pilots radar vectors from that point.
Avionics: Were there charting issues?
Townsend: There were two main design issues with the procedures. The arrivals all had runway transitions to every runway. So the charts looked like a pile of spaghetti because of all the routes in this small area. It was somewhat difficult to read in the beginning. To address this problem FAA decided to remove all the runway transitions. Today the procedures end at a point and then from that point pilots are vectored to the runway. The procedures take aircraft very close to the predominant landing runway. In fact, three of the four RNAV arrivals end at the initial approach fix for ILS to runway 25L. The benefit is, if you're using 25L, you essentially have a procedure that takes you all the way to the runway.
FAA fully intends to bring back runway transitions, now that we have better guidelines and criteria in place for it. In fact, Vegas is anxious to incorporate some runway transitions that would help them under various landing configurations.
Avionics: Were there other issues?
Townsend: There were issues relating to departures. There was a particular waypoint on a departure procedure that had an approximately 119-degree course change, so it was a pretty significant course change for an aircraft. It was a legal design and the aircraft did what they were supposed to do--fly by as closely as possible to the waypoint. But because it is such a sharp course change, an aircraft requires a certain amount of turn anticipation. It has to start the turn early enough so it can get as close as possible to the waypoint.
We discovered that, based on the different types of equipage on aircraft--the flight management systems [FMS]--and the angle of bank that was programmed into them, airplanes would start to turn to a waypoint at different distances [from it]. An Airbus aircraft might start a turn 3 or 5 miles out, whereas a G5 might start a turn 7 miles from the waypoint. What happens is that the G5 is now cutting in front of the Airbus. This created a potential loss of separation and caused a great deal of angst with controllers. Their initial reaction was to blame it on the procedure for not providing predictable and repeatable flight tracks. The procedures worked, but airplanes flew them differently.
Avionics: What was the solution?
Townsend: You soften the turn--you add a couple of waypoints to give the aircraft better guidance to follow these waypoints. This reduced the distance for the turn and made it more consistent for all of the airplanes. That's been a very important lesson learned for some of the other airports that have developed procedures since then. We discovered that if a turn of more than about 60 degrees is needed, turn anticipation will begin to vary at the waypoint. So air traffic managers need to determine if that's OK: for a particular waypoint is it OK for these airplanes to have inconsistent flight tracks associated with them? If it's not acceptable, then designers need to add waypoints for better guidance.
Avionics: So equipage is a big issue?
Townsend: We're discovering a lot of other issues with equipage--the automation on the aircraft. Because we need consistency with performance-based procedures, the airplanes essentially need to perform the same. The way the different manufacturers build their boxes affects the way airplanes fly certain leg types and procedures overall. FAA is beginning to work more closely with the manufacturers so that this can be corrected and we don't continue to run into these variances. At least 80 percent of all U.S. aircraft out there are capable of flying RNAV SIDS and STARS procedures. In the initial Vegas RNAV procedures implementation, they had just under 80 percent participation. Today it's around 95 percent. Conducting RNAV procedures requires strict adherence to lateral tracks in order to provide the consistency that ATC requires.
Avionics: How did you monitor the redesigned procedures?
Townsend: We also included an RNAV post-implementation team. It included myself, other airline representatives, union representatives from other airline groups, officials from the FAA's RNAV/RNP program office, flight inspectors and representatives from the TRACON. We were in a back training room, watching on the radar scopes. Any time we observed a glitch, we acted upon it. It was not a place to get anybody in trouble. It was to learn. For example, if a pilot on a departure did not comply appropriately with a vertical restriction, the airline representative immediately got on the phone and started a communications chain to make contact with the captain, when he reached the destination, to discuss what occurred. These interviews helped us tremendously with how pilots were interpreting the procedure and what the probable cause was that may have led to the misinterpretation.
Avionics: What problems were occurring?
Townsend: We underestimated the training that was needed for pilots and controllers. Flight management systems have been in use by air carrier airplanes for several years in the en route environment and with very simple arrival and departure procedures coded in the FMC databases. As we brought RNAV into the terminal realm, operational issues arose. There was a lack of effective training for the controllers and the pilots of most air carriers. Training for general aviation, including business aircraft pilots, was nonexistent. Certain procedure designs resulted in difficulties, as well (e.g., large turns, waypoints placed too closely together).
Air-side issues included improper programming of the flight management computer, failure to verify FMS inputs, misunderstanding of charted procedures, misunderstanding of certain ATC [air traffic control] clearances and improper execution of the procedures. We had several instances in Las Vegas, for example, where pilots were not selecting the en route transitions on departure. With this particular airline, the pilots were not following all the steps appropriately and then going back and verifying what they had programmed. That airline increased training and heightened awareness of this problem. And to my knowledge this rarely if ever occurs today at Las Vegas.
Errors will continue to be made in the NAS. Pilots and controllers are human and susceptible to errors. Radar provides a means for the controller to monitor the pilot's compliance with procedures. Even though we've automated or proceduralized a lot of the instructions that air traffic control would give, the controller retains the responsibility to control traffic and function as a "traffic manager."
Avionics: What about the ground side?
Townsend: Ground-side issues were the controllers' reluctance to adopt a new way of conducting operations, lack of confidence in pilots flying the procedures with predictability and reliability, and lack of knowledge on how the pilot interfaces with the FMS in terms of what the system can and cannot do.
Avionics: What were the solutions?
Townsend: Many carriers improved their training programs and placed emphasis on deficient areas. Controllers were encouraged to keep the airplanes on the procedures and alter the procedure only if necessary for safety or operational necessity. Charting was simplified on certain procedures. Data was collected and shared with the entire team in order to identify problems and fix them. Communications were open and frequent. Design issues were identified and corrected. Aircraft qualifications were revised. Procedures and phraseology were improved and clarified. LAS determined failure wasn't an option and worked towards a process of "evolution" instead of a "revolution." At Las Vegas everybody worked collectively to agree. That was the amazing part of the process--everybody was involved in it.
Avionics: Are there quantifiable benefits?
Townsend: From an operator standpoint, the fiscal benefits are important to the struggling airline industry bottom line. Since the procedures take advantage of space-based navigation and are not confined to the restrictions of the legacy nav infrastructure, distances are reduced, saving time and fuel. The vertical profiles that provide near flight-idle descents until close proximity to the airport not only save gas, but are also environmentally friendlier by reducing noise. Airlines are also saving money on structural components. The use of speed brakes and early extension of flaps and landing gear are less likely to occur, which reduce excess wear on the components.
More specifically, Las Vegas made some modifications to its departure procedures in March of 2005 to better hit the noise target area. With this redesign they reduced miles flown on all but one departure procedure. Nearly 7 miles were eliminated from one procedure. One airline estimated the one reduction would save them $500,000 a year in fuel costs, based on the number of departures they operate. And that was back in early 2005 before the astonishing rise in fuel costs.
From a pilot, controller and safety point of view, the procedures reduce workloads and reduce communications. As our nation's air traffic continues to increase beyond pre 9-11 levels, frequency congestion and read-back/hear-back errors are more of an issue. The RNAV procedures reduce communications by 30 to 60 percent. With fewer communications we don't have blocked calls. So the communications flow more efficiently.
Avionics: Is Las Vegas building on these procedures?
Townsend: Yes. They are ready to modify the procedures to take full advantage of runway transitions. They also are planning a right turnout procedure for departures. To give you a quick history, before the Four Corner Posts project, Las Vegas had departures that were both left and right turnouts. At the time of the redesign, it was determined that they would have all left turnouts, so everything moves in a counterclockwise direction out of Las Vegas. From an airline standpoint, if I'm going to a destination that's more to the north or northeast, it adds mileage and fuel burn. So adding a right turnoff procedure will offload some of this traffic, increase capacity and be a huge savings in operating cost.
Also Clark County, Nevada, is building a new airport scheduled for opening in 2017. The FAA and stakeholders already are working on procedures for this airport. Since McCarran airport is only about 30 miles away, a redesign of existing procedures will be necessary.
Avionics: Have other airports learned from the Las Vegas experience?
Townsend: In Atlanta, Delta Air Lines was the lead operator, and their procedures design expert had worked with us during the fixes for Las Vegas. He incorporated many of the lessons learned in the original procedure designs for Atlanta. And as a result, they had a very successful implementation. It was pointed out [to procedure developers] at Dallas--on their parallel departures--that in their original design they would probably encounter some of the same turn anticipation issues at specific waypoints that we experienced in Las Vegas. After some implementation issues with the original design, the departure procedures were modified to rectify the problem.
The FAA, along with the aviation community, has captured many of the lessons learned at LAS. FAA established an SMS [safety management system] program office within the last year. In fact, the first procedure that underwent safety risk analysis was the right turnout procedure for Vegas. SRM [safety risk management] is a total change in the way of doing business. The process enables us to identify and manage risk. It permits us to review past experiences and address known hazards at the same time we can look ahead and effectively apply safety risk management. SRM will become integral to design and implementation processes of performance-based procedures. Hazards can be identified and risks analyzed prior to implementation. This affords the opportunity to identify mitigation options. This process should help eliminate "band-aid" fixes that have been applied for previous mitigations and should ensure smoother implementations.
Avionics: Is the supply of RNAV procedures keeping up with demand?
Townsend: Currently, it takes a minimum of 200 days from the time an RNAV kickoff meeting is held until a procedure is published. A great deal of coordination is required, which makes it difficult to implement procedures as expeditiously as one may desire.
Avionics: Is the demand such that private procedure development is appropriate?
Townsend: The FAA plans no fewer than 155 RNAV procedures and routes for FY06 and 167 for FY07. I believe this number will continue to escalate as RNAV is more widely accepted. "Inquiring minds" want to know when the resources of the National Flight Procedures Group will be stretched to the limit, not only with new procedures, but with maintaining existing ones. Our good neighbors to the north in Canada have been capitalizing on the services of third-party developers with a measurable amount of success. The FAA administrator has stated an interest in exploring this concept for the U.S. I agree it should be considered, but let's make certain it is conducted appropriately.
Avionics: ALPA continues to play a role?
Townsend: ALPA is deeply involved with the technical criteria aspects. We have representatives on the PARC [Performance-based operations Aviation Rulemaking Committee] and various industry working groups. ALPA assists in all stages of RNAV and RNP implementation, including development of procedure design criteria, aircraft airworthiness standards, operational guidance, simulator and other trials, and pilot education.
What I'm trying to implement now is to get our local ALPA ALRs [airport liaison representatives] and RSCs [regional safety coordinators] involved when procedures are under development, so we're able to provide the line pilot perspective and defend the ALPA safety case for these procedures.
Avionics: What's the lead carrier model? Is it sometimes misunderstood?
Townsend: The FAA has applied the "18 Step Process," which provides guidelines for implementing terminal RNAV procedures. It requires the designation of a lead operator to provide flight technical expertise. We have discovered that FAA flight checks do not always mirror the way air carrier operators fly the procedures. Therefore, it's imperative that procedures are evaluated in simulators with various types of aircraft. The lead carrier is a coordinator between other carriers and provides technical expertise on the operational end and simulator evaluation of the procedures to make sure they are designed so that they can fly them. The lead operator should encourage operators with other aircraft types to check the new procedures in their simulators. In certain cases, there tends to be a lack of emphasis on the lead carrier's role and responsibilities.
Avionics: How can RNAV design and implementation be improved?
Townsend: Three words--collaborate, communicate and educate. It takes many different divisions of the FAA to work together, as well as closely working with the customers. Communicating by sharing data, ideas and information is central to success. Educating is probably the most challenging and significant piece of the process. Procedures designers must be made aware of the operational needs of the customer, and the customer should understand the air traffic demands for the facility. Equally important is proper training for controllers and pilots. The need to redesign the surrounding airspace to facilitate designing efficient procedures also continues to be a factor. Developing RNAV procedures around inefficient airspace can be detrimental. Remaining focused on long term goals in spite of short term problems is integral to RNAV's success.