Proceedings of the 05 Systems and Information Engineering Design Symposium Ellen J. Bass, ed. RUNWAY OPERATIONAL QUALITY ASSURANCE (ROQA) SYSTEM Janet Geldermann Yusuf Mohamed Robert Smith Mahmoud Yessad Faculty Advisor: Dr. George L. Donohue Department of Systems Engineering and Operations Research 4400 University Drive George Mason University Fairfax VA 2 ABSTRACT Bottlenecks are developing in the air traffic system that controls airline traffic flows. Runways are a limiting factor at airports where consumer demand continues to increase and/or runway construction has been limited by land constraints. One approach to raise system capacity and to lower bottlenecks is to reduce the time separations between aircraft during the approach phases of flight. Such changes should be quantified to document the safety impacts of separation reductions. A team of Systems Engineering students from George Mason University has developed the initial design for a runway monitoring system, the Runway Operational Quality Assurance (ROQA) System. Similar to the FAA Flight Operational Quality Assurance (FOQA) system that analyzes individual aircraft operational flight data, ROQA is designed to provide operational performance measurements of approach runway operations. ROQA data is analyzed for operational parameter exceedences and statistical distributions are calculated to determine collision risk. 1 INTRODUCTION 1.1 System Background One billion air passengers a year are expected to be traveling by US commercial airlines within the next decade. [04 FAA forecast] This is a forty-five percent increase in passengers by 15. In the first eight months of 04, US airlines carried 7.5% more passengers than in the same period the previous year and flew 3.9 % more flights, not including air cargo flights. The number of international passengers was up 9% [04 Administrator s Fact Book] Commercial operations, at towered airports, in 03 numbered 28.3 million with forecasted operations by to number around 32.7 million and 39.1 million in. According to the Airports Council International - North America (ACI-NA), approximately twenty-two runways have been or will be completed by the end of 05 just to accommodate year 00-demand levels. However, new runway construction at many of the major airports has been capped. New accommodations must be made to meet the forecasted demands at airports in the National Air Space (NAS). As demand for airport capacity increases safety must also be increased. Current worldwide accident rates are estimated by a National Aerospace Laboratory (NLR), Amsterdam, to be approximately 1 X -9 or one fatality in one billion flight hours [Blom et al 01]. Merely maintaining the current safety rates while increasing capacity may raise the number of accidents. One of the Federal Aviation Administration Air Traffic Organization (FAA-ATO) Safety Assurance Office goals is to reduce air carrier fatality rates by eighty percent by 07 from 1996 numbers.[faa Strategic Plan] Fifty-one percent of commercial aviation accidents occur in the final stages of flight, either on final approach, landing and/or on airport runways. [Boeing Statistical Summary] Precursors are defined as events that could lead to an aircraft accident or collision. It was recently announced the year 04 was the safest for air transportation since the 1940 s. Therefore, during this period of time baseline metrics of National Airspace System (NAS) runway operations should be documented. Operations should continue to be monitored as the NAS increases its capacity to study and monitor the impact upon safety. 1.2 Problem Statement The FAA - ATO Safety Assurance Office is required to ensure the safety of the NAS. In order to measure system safety, data collection on the collision precursors should occur and then be transformed into stochastic statistics for determination of the safety runway operations. The current statistical standards used by the FAA-ATO Safety Assurance office do not adequately reflect runway operations.
Stochastic metrics must be used to baseline the current runway operations. Performance metrics are used not only to quantify operations but also to improve system operations. Secondly, some airports are already operating at reduced separation standards and precursors are not consistently being reported by flight crews and air traffic controllers. Additionally, in order for the FAA to reduce the required separation standards to allow increased capacity at major airports, actual metrics must be recorded and analyzed. To serve this purpose the ROQA team has been tasked with developing the initial design of a runway monitoring system. 1.3 Separation Regulations & Precursor Descriptions The FAA regulations have determined separation distances and times for aircraft depending upon the weight classifications of leading and trailing aircraft. Aircraft weight classes are defined as: Heavy capable of takeoff weights of greater than 255,000 pounds B757 - specialized category for large wake vortex size Large - capable of takeoff weights of greater than 41,000 pounds but not more than 255,000 pounds Small capable of takeoff weights of not more than 41,000 pounds Current FAA separations for the above weight classes are found in Table 1. The table includes the nautical mile (nm) distance between lead and trailing aircraft and the conversion to time based on average aircraft velocities. Wake Vortex encounters, which are discussed later in this paper, are the main reason these separations exist. Table 1: Aircraft Weight Class Separation Standards Matrix Leader Trailer Heavy (1) B757 (2) Large (3) Heavy (1) B757 (2) Large (3) Small (4) 99 sec ( 4 nm) 99 sec 62 sec 129 sec (5 nm) 3 sec 129 sec (5 nm) 3 sec 166 sec (6 nm) 138 sec (5 nm) 111 sec time an aircraft crosses the threshold to the time the said aircraft s tail fully exits the runway. Figure 1: IAT & ROT Diagram 2.1 Inter-Arrival Times and Distances Inter-arrival time (IAT) is defined as the time separation between a lead and trailing aircraft crossing the same runway threshold. Inter-arrival distances are calculated based upon the inter-arrival times between the lead and trailing aircraft and the aircraft velocities and deceleration rates. 2.2 Inter-Departure Times and Distances Inter-departure times are defined as the time between two aircraft taking off the same runway. Take off is considered to be wheels up. Inter-arrival distance is the distance between two departing aircrafts on the same runway. Again, times and distance vary depending upon the velocities and accelerations of aircraft. 2.3 Runway Occupancy Time There are two types of runway occupancy times (ROT). For an arriving aircraft, it is defined by the Airport Capacity Modeling Task Force as the time interval between crossing the threshold and the aircraft s tail vacating the runway. Departure ROT is the time interval between crossing the holding stop bar and the main gear lifting off the runway. [Eurocontrol, January 03] Both of these types will be referred to only as runway occupancy times for this paper. 3 PRECURSORS AND COLLISION PROBABILITIES Small (4) 62 sec ( 2.5 nm) 69 sec 3.1 Simultaneous Runway Occupancy (SRO) 2 RUNWAY DATA MEASUREMENTS Figure 1, IAT & ROT Diagram, shows which portion of the approach and landing phases of a flight addresses the inter-arrival times (IAT) and runway occupancy times (ROT). IAT is measure from the time the lead aircraft crosses the runway threshold to the time the trailing aircraft crosses the threshold. ROT is measured from the Simultaneous runway occupancy is the event of two or more aircraft occupying the same runway at the same time. Typically, when an aircraft is given clearance to land or take off it has ownership of the specified runway. Simultaneous runway occupancies can occur if an aircraft has crossed the runway threshold but a previously landed aircraft has not completely cleared the same runway. 135
3.2 Wake Vortex Encounters Wake Vortex Separations are based on wake vortex encounters. An aircraft s wings create wake vortices during flight. A wing's lift is due to the difference between the higher pressure acting on the wing's bottom surface and by the lower pressure acting on its top surface. As a byproduct of this differential in pressure, the airflow near the wingtips tends to curl around the tips, moving from the higher-pressure region below the wing to the lower pressure above the wing. This air movement creates a rotating vortex downstream from each of the wingtips. [Rannoch 04] Figure 2, Wake vortex diagram, shows wake vortices created at the wingtips of an aircraft. The size of these vortices is dependent upon the aircraft weight class, speed, wingspan, and weather conditions. Larger aircraft can create vortices of up to 0 feet in diameter. The circulation at the center of the vortex can be upwards of 0 feet per second. Once an aircraft is on the ground the wing no longer creates lift, therefore wake vortices are no longer being produced and existing ones will dissipate in approximately three minutes. If there is a breeze or wind in the airport area, the vortex will dissipate more quickly. [Assessment of Wake Vortex Separation] to perform data collection, analysis, and monitoring of the operations for assessing opportunities for improved performance of the NAS. ROQA is a decision support system for the purpose of collecting the operational data of aircraft operating on approach, landing, and departure at any controlled airport within the NAS. ROQA will receive and record sensor inputs that will be calculated into aircraft positions and collisions precursors that are further calculated into collision risk probabilities. The precursor information and the safety risk probabilities are then provided to the FAA-ATO Safety Assurance Office. The office can use the data for decision-making purposes specifically related to reducing aircraft separation standards and thus increasing airport capacities. Based upon the highly successful FAA Flight Operational Quality Assurance (FOQA) program, ROQA will record and provide operational performance measurements and metrics to the FAA-ATO Safety Assurance Office. However, unlike FOQA data, which is collected and analyzed by the airlines safety office before transmission to the FAA, ROQA will be run by the FAA-ATO for the FAA-ATO. This will provide objective accurate data and information for determining the safety situation of any runway where ROQA has been deployed. The ROQA System design diagram, Figure 3, ROQA System Interfaces, is based upon Department Of Defense Architecture Framework, OverView-1 (DODAF OV-1) requirements. Aircraft transponders transmit signals, which are received by airport multilateration sensor equipment placed at strategic sections of the runway surface. (Multilateration equipment has a higher update rate and is more accurate than the surface surveillance radar.) The multilateration equipment transmits aircraft position, weight class, velocities to ROQA. Figure 2: Wake vortex diagram Rannoch Corporation A trailing aircraft, landing or taking off too closely to the leading aircraft may have an adverse encounter with a wake vortex produced by the lead aircraft. The circular motion of the wake vortex could disrupt the smooth airflow over the trailing aircraft s wings adversely impacting control. During landing or departure sequences, an aircraft often does not have enough altitude to recover and the results can be disastrous. 4 SYSTEM DESIGN 4.1 System Overview The user of the Runway Operational Quality Assurance System is the Federal Aviation Administration (FAA) Air Traffic Organization (ATO) Safety Assurance Office. As a performance-based organization, the organization s role is Figure 3: ROQA System Interfaces 136
The data received by ROQA is processed or scrubbed to remove outliers such as runway occupancy times of zero seconds. Runway occupancy times (ROT) and inter-arrival times (IAT) means and standard deviations are calculated, recorded and transformed into the Statistical Measures of Performance which are then displayed via computer interface at air traffic control towers and at the FAA-ATO Safety Assurance Office. The means and standard deviations are used to compute the safety risk metrics. The Runway Occupancy Times are used to calculate the probability of Simultaneous Runway Occupancies (SRO) based upon algorithms developed by Dr. John Shortle and Richard Xie at George Mason University (GMU). IAT mean and standard deviations are the basis for calculating the probability of a wake vortex encounter. Algorithms for computer Wake Vortex Encounter (WVE) probabilities are currently under study at GMU by Shortle, Xie, and Peter Choroba of Eurocontrol Experimental Centre in France [Xie 04]. Collision probabilities will be displayed for the FAA-ATO in a similar manner as the statistical measurements. Figure 4, ROQA System Measurements, displays the relationships between the actual data measurements, the statistical performance measurements, and the safety risk metrics that are necessary for the FAA-ATO Safety Assurance Office to monitor the performance of runway operations. round. The system will operate within a radius of five nautical miles (nm) of the airport with a concentration on final approach and landing sequences and departures. Only approach and landing will be addressed in this initial design of the system. 5.2 Value Hierarchy Figure 5, ROQA Value Hierarchy, displays a second level value hierarchy. The weights utilized in the value hierarchy were estimated by the ROQA team based upon the purpose and objective of the ROQA System. They were then reviewed with the project technical advisor and industry advisor for relevance and correctness. The two main value categories of the hierarchy are system performance and cost. The performance category is the more critical of the two and is given the greatest weight consideration. Performance is further broken down into Accuracy, Availability, Reliability, Maintainability, and Usability. Accuracy is weighted significantly higher than other system characteristics. As a result the ROQA system must be constituted of high precision components that will accurately monitor runway traffic. Actual Data Measured Runway Threshold to Runway Exits Statistical Measure of Performance Runway Occupancy Time (ROT) - pdf Safety Risk Metrics P(Simultaneous Runway Occupancy SRO) Algorithms from Xie/Shortle Distance between aircraft pairs @ runway threshold (Time mix) Inter-arrival Distance pdf Inter-arrival Times pdf Aircraft Fleet Mix Markov probability model P(Wake Vortex Encounter) Algorithms from Xie/Shortle P(Separation Violation) Algorithms - TBD Figure 4: ROQA System Measurements 5 OBJECTIVE & SCOPE As a decision support tool, ROQA will present information for FAA-ATO Safety Assurance Office to monitor the safety of NAS runway operations and provide a basis for decision-making activities. The approach to increase the capacity of airports in the NAS could benefit from this system. Some airports are currently operating below prescribed aircraft separation standards and ROQA can reveal whether this has become standard operating procedure for airports especially during push times or in times of increased demand. For example, Atlanta s Hartsfield Airport already has aircraft separations reduced from the FAA regulation standards of three miles to a 2.5-mile aircraft separation. 5.1 Scope ROQA is proposed to be operational at any controlled or towered airport within the NAS, in all weather, 24/7 year Figure 5: ROQA Value Hierarchy Availability represents the probability that the system will be operating as specified by the stakeholders and was given a weight of 0.25, the second highest weighting. In order to provide accurate up-to-date statistical information the system should run consistently. Reliability, maintainability, and usability are not significant factors of the system. Security and safety are irrelevant characteristics, as the system does not affect the security of the airport or the safety of runway operations. No one will be adversely affected by a malfunction of the system. 137
While cost is always, an important factor in any system, it is not as important as the system performance and was given a lower weight of 0.40. Cost is broken down into four categories. Implementation Costs are the initial costs of system development and weighted as. as are the Return on Investment, and maintenance costs. These categories are lower in importance compared with the Cost of Improved Performance. Cost to insure the stakeholders obtain the performance they desire received a significant weight of.70. 5.3 Functional Architecture The ROQA system project proposes the initial design of a runway operations monitoring system for monitoring safety. Figure 6, External Systems interface with ROQA Functions, provides a functional breakdown of the ROQA system within the boundaries of the center shaded box. option for the Runway Statistical Measure of Performance Display. The primary property of the vertical meters is the numerical range from low to high. The colors reflect the distribution of the FAA separation standards of an airport runway operations based upon benchmark figures from a NAS airport. As typical in the US, the green color represents a system that is operating normally, yellow - caution and red dangerous conditions. The current operating mean of the measurement is shown by the large arrow. The narrower arrows to either side characterize the standard deviation of the metric. Finally, the mean and standard deviation are presented numerically below the statistical meter. 5XQZD\6WDWLVWLFDO0HDVXUHVRI3HUIRUPDQFH 5XQZD\ 2FFXSDQF\7LPH,QWHU$UULYDO 7LPH Multilateration Sensor Equipment Receive Inputs, Provide Preprocessing of Data Provide Statistical Measurement Calculations Provide Safety Risk Metric Calculations FAA-ATO Safety Assurance Office $YHUDJH5XQZD\ $YHUDJH,QWHU $UULYDO 2FFXSDQF\7LPHLQ 7LPHLQ6HFRQGV 6HFRQGV XQZD\2FFXSDQF\ QWHU $UULYDO7LPH 7LPH6WDQGDUG 6WDQGDUG'HYLDWLRQLQ 'HYLDWLRQLQ6HFRQGV 6HFRQGV FAA Separation Standards Provide User Interface ATC Tower Figure 7: User interface 5.4 Design Alternatives ROQA Figure 6: External Systems interface with ROQA Functions Receive Inputs, Provide Preprocessing of Data will receive the inputs from the surface surveillance equipment at an airport. Multilateration equipment provides the most accurate and timely data, although ROQA could also work with traditional surface surveillance equipment. The positional data provided by the multilateration equipment will be scrubbed to remove outliers in the data such as a zero velocity, and calculate runway occupancy and inter-arrival times. Provide Statistical Measurement Calculations will compute the mean and standard deviations of ROT and IAT. Provide Safety Risk Metrics receives the statistical measurements and computes the collision probabilities. Both the statistical measurement calculations and the safety risk metrics are provided to the FAA-ATO via the Provide User Interface function of the ROQA System. This function converts ROQA statistical measurement and metric calculations into graphical information displays for the users. Figure 7, the User Interface, shows one possible No systems similar to ROQA are currently in production or in use anywhere. The system can be produced using Commercial Off the Shelf (COTS) hardware and software that can be customized for ROQA purposes. This provides a relatively inexpensive means to build a prototype for system testing. The system can also include functionality that can switch the system from measuring operations under the current FAA operating standards to stochastic standards. Stochastic standards would allow the FAA to view the operations at reduced standards against the current operations before officially changing the standards. Such testing can provide data and information to the FAA for decision analysis. 6 SIMULATION 6.1 Approach The ROQA simulation is used to demonstrate the experimental outcomes of various time separations of aircraft on 138
approach and landing, and the impact on runway operational safety. Figure 8, Simulation Logic Model, displays the logic of the simulation. The development software used to create the software was Arena, version 7.01 by Rockwell Software, Inc. Simulation aircraft time sequences are based on historical data and field observations of landing and runway operations of airports found in the NAS. Various simulation scenarios include varying the distribution of interarrival times and the use of Markov chains for creating aircraft type mixtures. These provide the input to the simulation. The outputs of the simulation provide statistical measures of performance of runway operations. Frequency 35 25 15 5 0 Empirical (LGA) 40 50 60 70 80 90 0 1 1 1 140 150 160 170 180 190 0 2 2 2 240 250 260 270 280 290 0 IAT (sec) Figure : Empirical Data Collected March 05 CONCLUSION Frequency 40 0 Figure 8: Simulation Logic Model Simulation of IAT 40 50 60 70 80 90 0 1 1 1 140 150 160 170 180 190 0 2 2 IAT (sec) Figure 9: Simulated IAT histogram Figures 9 and show histograms comparing sample simulated data versus empirical data for aircraft interarrival times. By varying aircraft approach speeds and runway occupancy times, based on actual data, the simulation will provide reliable output data for determining safe approach and landing operations. The demand for air transportation is the United States is rapidly increasing while capacity at major airports the NAS are approaching maximum limits under the current FAA separation standards. Runway capacities and wake vortex encounters on approach are two of the limiting factors to increasing airport capacities. ROQA, as a decision support system will provide initial baseline performance metrics and continue to measure runway operations. The ability to monitor runway operations based upon current FAA separation standards should provide the recognizable stochastic nature of operations. ROQA will provide the ability to monitor runway operation in the event of aircraft arriving closer than separation regulation distances. Resulting information may be utilized to change separation specifications and allow more aircraft to land on NAS runways. Unlike the FOQA system, ROQA should be mandatory; run by the FAA-ATO for the FAA-ATO and provide objective data and information for decision-making analysis. The data will be sanitized due to air traffic control and airline liability concerns. Finally, ROQA will supply accurate statistical representation of runway operation of the Aviation Safety Statistic Handbook which is published on a monthly and yearly basis by the FAA. ACKNOWLEDGEMENTS The ROQA team would like to acknowledge Dr. Benjamin Levy and the SENSIS Corporation of DeWitt, New York for their sponsorship and advice developing this project. We would also like to acknowledge Dr. Lance Sherry, of GMU for his guidance on air traffic control, and runway concerns throughout the project life. Finally, we would like to recognize Dr. John Shortle and Richard Xie of GMU for their advice and willingness to share their runway landing safety analysis simulation used in conjunction with our own simulation. 139
REFERENCES Administrator s Fact Book. November 04. <http://www.atctraining.faa.gov/fact book> Airport Capacity Modeling Task Force Glossary. Eurocontrol January 03. Assessment of Wake Vortex Separation Distances. October 04. National Aerospace Laboratory, NLR ( Speijker, van Baren), Center for Air Tranportation Systems Research CATSR (Sherry, Shortle), NASA Langley Research Center (Rico-Cusi). Blom, H.A.P., et al. Accident Risk Assessment for Advanced Air Traffic Management. Air Transportation Systems Engineering, 01. Eds. Can the wake vortex be eliminated? Retrieved November 25, 04. Rannoch Corporation. <http://www.rannoch.com/wakeaerf.htm l> Federal Aviation Administration (FAA) Aerospace Forecasts, Fiscal Years 05 16. Department of Transportation. <http://www.api.faa.gov/forecast05/f orecast%for%05.htm> [Accessed March 19, 05] Wieland, Frederick. March 05. Estimating the Operational Capacity of the National Airspace System. AEgis Technologies Group. Alexandria, VA. Xie, Yue, John Shortle, and George Donohue. October 03. Runway Landing Safety Analysis, A Case Study of Atlanta Hartsfield Airport. Available online via <http://mason.gmu.edu/~yxie/dasc%2 003_Landing%safety.pdf> [Accessed October 04]. Xie, Yue, John Shortle, and Peter Choroba. 05. Quantitatively Estimating Wake Vortex Safety Using P2P Model. Available online via <http://mason.gmu.edu/~yxie/6th_atm. pdf> ROBERT SMITH is currently a senior in Systems Engineering at George Mason University specializing in Telecommunications, Computer Infrastructure Development. After graduation, he plans to become an independent contractor for technology firms and government agencies. He can be reached at <rsmithd@gmu.edu> MAHMOUD YESSAD is a senior in Systems Engineering at George Mason University specializing in system modeling and performance. After graduation, he plans to obtain employment with a systems engineering firm to apply and expand his knowledge of the systems engineering field. He can be reached at <myessad1@gmu.edu> AUTHOR BIOGRAPHIES JANET GELDERMANN is currently a senior in System Engineering at George Mason University. Her specialization area focus has been on modeling and decision support systems. After graduation, she will pursue a career in the aerospace industry while working on a master s degree in Information Security and Assurance. She can be reach via email at <jgelderm@gmu.edu> YUSUF MOHAMED is currently a senior in Systems Engineering at George Mason University specializing in networking. After graduation, he plans to continue working full time at his current job at Ilumin Software Services. He can be reached at <ymohamed1@gmu.edu> 140