Evaluating Airfield Capacity (2012)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

31 Existing Airfield Capacity Evaluation Tools The purpose of this chapter is to identify the various levels of current modeling sophistication for analyzing airfield capacity and to describe typical examples and features of each level. Although they are out-of-date, FAA’s AC 150/5060-5, Airport Capacity and Delay (the AC), and Airfield Capacity Model (the ACM) provide a useful starting point for discussing the levels of modeling sophistication. The four levels implied in the AC are as follows: 1. Table Lookup, as illustrated in Chapter 2, “Capacity and Delay Calculations for Long Range Planning,” in the AC 2. Charts, Nomographs, and Spreadsheets, as illustrated in Chapter 3, “Airport Capacity and Aircraft Delay Calculations,” in the AC and new spreadsheets presented in Chapter 4 of this guidebook 3. Analytical Capacity and Delay Models, as described in Chapter 5, “Computer Programs for Airport Capacity and Aircraft Delay,” of the AC these models are computer programs for calculating airfield capacity and aircraft delay as described in Chapter 3 of the AC. 4. Airfield Simulation Models, as described in Chapter 5 of the AC, are computer programs that include models such as the FAA Airport and Airspace Simulation Model (SIMMOD), the FAA Airfield Delay Simulation Model (ADSIM), and proprietary software. ACRP Report 79 focuses on the first three levels of modeling sophistication plus a fifth level identified by this project’s research and illustrated by the MITRE Corporation’s runwaySimu- lator. The runwaySimulator differs from earlier models like SIMMOD, ADSIM, and the Total Airspace and Airport Modeler (TAAM)—and even from runways-only simulation models such as FAA’s Runway Delay Simulation Model (RDSIM)—in that it has been designed to produce estimates of airfield capacities (maximum sustainable throughputs) rather than aircraft delays, although it can be used to estimate both. As such, the MITRE model is likely to be less intensive in relation to data and workload and therefore will likely require fewer resources to apply than the conventional airfield simulation models. Table 3-1 presents information about the five levels of modeling sophistication discussed in this guidebook. The range of modeling analyses and techniques presented in Table 3-1 should provide airport operators and airfield capacity analysts with appropriate levels of sophistication to address the wide range of capacity analyses issues they are likely to encounter. Identification of a new level of sophistication (airfield-throughput simulation models) fills in an important gap between the analytical models and the simulation models. The five levels of modeling sophistication are further described in the remaining sections of this chapter, with respect to the following characteristics: • Applications: Typical capacity issues that can be analyzed with the level of modeling sophis- tication (i.e., what is this level of modeling sophistication best suited to do?) C h a p t e r 3

32 evaluating airfield Capacity • Modeling assumptions: Fixed, built-in assumptions within the level of modeling sophistica- tion that cannot be altered through input parameters (i.e., what assumptions are hardwired into the model and cannot be changed?) • Data requirements: Required data inputs for the level of sophistication (i.e., what are the data requirements for defining the inputs and validating the outputs of the modeling?) • Time and cost requirements: Typical time and cost required for a capacity analysis using that level of modeling sophistication (i.e., what are the approximate resource requirements in terms of both elapsed time and cost?) • Model availability: Cost and process for acquiring the model (i.e., is the model publicly avail- able, and what is the cost associated with acquiring the model?) • Operator skill and training required: The specific background, knowledge, and training required for the user to effectively use the tools under each level of sophistication Level Descrip�on Examples Sample Applica�ons A�ributes/Limita�ons Data Requirements 1 Table lookup Chapter 2 of the AC, new lookup table Statewide system plans, airport master plans where airfield capacity is not an issue, and small airport master plans Runways only, simplified airfields, small airports, default assump�ons only Minimal, requiring only an overview of airport runway configura�on and aircra� fleet mix 2 Charts, nomographs, and spreadsheets Chapter 3 of the AC, new spreadsheet model Statewide system plans, airport master plans where airfield capacity is not an issue, and small airport master plans Runways only, moderate size airports, less complex airfields, some flexibility in inputs Minor, requiring airport runway configura�on, aircra� fleet mix, exit loca�ons, and percentage of arrivals 3 Analy�cal capacity models Airfield Capacity Model Specialized airfield capacity studies, airport master planning studies, regional airport system planning Runways only, moderate airfield complexity, taxiways and airspace considered implicitly, flexible input assump�ons More demanding, including aircra� fleet mix, aircra� final approach speeds, aircra� separa�ons, and air traffic control (ATC) rules 4 Airfield capacity simula�on models runwaySimulator, Flexible Airport Simula�on (FLAPS) Capacity planning of complex airfields or regional airfield/ airspace systems Runways only, complex airfields and airspace, flexible assump�ons More detailed input data than Level 3 models, including close-in arrival and departure flight track geometries and aircra� fleet mix by runway 5 Aircra� delay simula�on models SIMMOD, ADSIM, TAAM Detailed planning of complex airfields or regional airfield/ airspace systems Runway, taxiways, aprons, gates, and/or airspace; complex airfields (e.g., runway crossings and airspace fix constraints), flexible input Greatest level of detail about aircra� flight schedule and airfield and airspace configura�ons, including taxiing routes and aircra� parking posi�ons Source: LeighFisher. Table 3-1. Proposed levels of modeling sophistication.

existing airfield Capacity evaluation tools 33 Level 1—Table Lookup This level of modeling sophistication refers to capacity analyses completed by table refer- ence, as exemplified in Chapter 2 of the AC and the prototype Airfield Capacity Spreadsheet Model described in Chapter 4 of this guidebook. The table reference mainly involves looking up an airport’s runway configuration along with certain other airport characteristics, most commonly fleet mix, to determine hourly runway capacity and annual service volume (ASV) in visual meteorological conditions (VMC) and instrument meteorological conditions (IMC) (Figure 3-1). Applications Table lookup methods typically consider runways only and are best used for high-level capac- ity analyses conducted as part of the creation of airport system plans or smaller airport master plans. The most suitable applications of this level of modeling sophistication are estimations of (1) existing hourly runway capacity and ASV, or (2) major capacity changes, such as those associ- ated with additional runways, new runway configurations, increased spacing between runways, or a significant change in aircraft fleet mix. Modeling Assumptions This level of modeling sophistication cannot capture or change any of the default assump- tions used in estimating the hourly capacity and ASV values presented in the tables. The AC provides capacity estimates for 19 runway configurations, representative of typical U.S. airports that have these configurations. However, if an airport’s runway configuration is not included, or if operational conditions differ from those assumed for the tables, this method cannot be used. For example, standard IMC/VMC separations are assumed in the capacity values, and these assumptions cannot be adjusted. In the AC, it has been assumed that the airport has no airspace limitations and that, if instrument flight rules (IFR) capacity is desired, at least one runway is equipped with an instrument landing system (ILS). In addition, these assumptions cannot be altered to reflect future technologies or changes in flight procedures. Also, it has been assumed in the AC that each runway has a full-length parallel taxiway. Tables that can be used to estimate Source: Federal Avia�on Administra�on. Figure 3-1. Portion of lookup table from the AC.

34 evaluating airfield Capacity the effects of a partial parallel taxiway or no parallel taxiway are not provided in this section of the AC, although a figure is provided in Chapter 4 (Figure 4-26) for this purpose. Data Requirements The data requirements for table lookups are minimal, requiring only airport runway configu- ration and percentage of large and heavy aircraft in the fleet mix. Aircraft fleet mix data typi- cally are readily available from airport records, OAG (formerly Official Airline Guide), or from FAA’s Enhanced Traffic Management System Counts (ETMSC). It is reasonable to assume that at smaller airports very few large and no heavy jet aircraft are in the fleet mix. Little or no opportu- nity or capability exists to improve the quality of the Level 1 capacity estimates with additional or improved data and assumptions. Time and Cost Requirements This method of capacity analysis requires a small investment commensurate with the fidelity of the capacity estimate obtained through these methods. Completing a capacity analysis with this level of sophistication can be expected to take about a day, and would likely cost less than $5,000. Model Availability Level 1 models are publicly available as part of the AC, which is available free of charge on the FAA website (www.faa.gov). Other Factors Chapter 2 of the AC was last updated in 1983. Despite the age of the method contained in the AC, it is still widely used in the United States and worldwide. However, the method does not allow for consideration of (1) changes that have occurred in ATC rules and procedures since 1983, or (2) variations in assumed ATC rules or future flight procedures and their effects on the capacity of different runway configurations. The data presented in the table lookup method could be updated to reflect current rules and procedures, but the methodology still could not be used to evaluate future rules and procedures, such as those that might be associated with NextGen technologies. Level of Operator Skill and Training Required Users of the table lookup method need to have a basic knowledge of airfield operations and the factors that affect airfield capacity, such as aircraft fleet mix, runway use configuration, and weather conditions. With such a background, little or no training is required for using the table lookup method. Summary of Limitations The main limitations of the currently available techniques at this level of sophistication are the unchangeable built-in assumptions for calculating the capacities and the limited number of runway use configurations available (currently 19 in the AC). The values of hourly and annual capacity provided in Chapter 2 of the AC are derived from typical capacities achieved at U.S. airports that have similar runway use configurations, reflecting ATC rules and proce- dures from 1983. However, when more precise capacities are required, or if the conditions at

existing airfield Capacity evaluation tools 35 the airport do not match the built-in assumptions, a higher level of modeling sophistication must be used. Nevertheless, the existing table lookup method is adequate for the purposes of calculating a quick, broad-brush, runway-focused capacity estimate for an airport with an airfield configura- tion available in the lookup table and with operating characteristics that follow the underlying assumptions. Otherwise, a higher level of modeling sophistication should be used. Level 2—Charts, Nomographs, and Spreadsheets This level of modeling sophistication refers to capacity analyses completed for a wide range of runway use configurations and operating alternatives through the use of charts and nomo- graphs, as presented in Chapter 3 of the AC. Level 2 modeling gives the user some additional flexibility in specifying operating conditions at the airport. To calculate hourly runway capacity, this method requires the selection of the best represen- tation of ceiling, visibility, and runway use configurations from a diagram of various runway use configurations. The user must then find the corresponding charts or nomographs for the selected case. Figure 3-2 illustrates the charts/nomographs method as applied to a particular runway use and weather condition. As shown in the figure, the capacity represented by the selected configu- ration is determined from the graph, relating mix index with hourly airfield capacity for different Source: Federal Avia�on Administra�on. Figure 3-2. Portion of chart/nomograph from the AC.

36 evaluating airfield Capacity percentages of arrivals. The capacity can then be further adjusted to account for (1) the percent- age of touch-and-goes (training flights) at the airport and (2) runway exit factors, through the application of various adjustment factors determined from the charts and information about the traffic at the airport and the number of runway exits and their locations. This method can be repeated for all runway use configurations and weather conditions at the airport. If desired, ASV can then be calculated by first taking the weighted average of the hourly capaci- ties over the year (which is computed using formulas specified in the AC), and then expand- ing that weighted average up to an annual number by multiplying by the ratio of average day, peak month (ADPM) operations to peak-hour operations, and the ratio of annual operations to ADPM operations. Thus, ASV is calculated using the following formula: ASV C D H,w= × × where Cw = the weighted average hourly capacity of the airfield, D = the ratio of annual to ADPM demand, and H = the ratio of ADPM demand to peak-hour demand. The D and H can be determined from airport records or publicly available sources on air traffic demand patterns. In cases where such demand data are not available, the recommended default values in the AC can be used. Applications This method for calculating airfield capacity is more refined than the table lookup method. However, Level 2 modeling is still considered best used for high-level capacity analyses conducted as part of the development of system plans or master plans because the charts and nomographs generalize the airport’s operating configuration and activity. The chart and nomograph method is generally considered most suitable for smaller airports and simpler runway configurations, and preliminary evaluations where limited resources are available. The most suitable applica- tions of this level of modeling sophistication are estimations of (1) individual capacity-related components, (2) hourly airfield capacity and ASV, or (3) major capacity changes, such as those associated with additional runways, new runway configurations, or increased spacing between runways. Modeling Assumptions This level of modeling sophistication cannot capture or reflect operational constraints or deviations from the assumptions used in developing the charts and nomographs included in the AC. The AC provides capacity estimates for 43 runway configuration and weather combina- tions, representative of typical U.S. airports. However, if an airport’s runway configuration is not included, or if specific operating constraints are not included, then this method cannot be used. Standard IFR/VFR separations are assumed in the capacity values, and these assumptions cannot be adjusted. It was assumed in the AC that the airport would have no airspace limitations and that the AC would not apply to situations such as the absence of an Airport Traffic Control Tower (ATCT), a parallel taxiway, or an ILS, except as provided in the previously mentioned Figure 4-26 of the AC. This section of the AC also would not apply to airports that have aircraft type restric- tions on certain runways. In addition, there is no possibility of altering these assumptions to reflect future technologies or changes in flight procedures. The new Prototype Airfield Capacity Spreadsheet Model developed in the research for ACRP Project 03-17 and described in Chapter 4 of this guidebook has been designed to overcome many of these limitations.

existing airfield Capacity evaluation tools 37 Data Requirements The data requirements for this method are minor, requiring airport runway configuration and airport geometry/layout information, specified percentages of large and heavy aircraft in the fleet mix, and exit locations and percentage of arrivals. Nearly all of this information is readily available from airport records, the OAG, or FAA’s ETMSC. At smaller airports, it is reasonable to assume that relatively few large and no heavy jet aircraft are in the aircraft fleet mix. Limited opportunity and ability exist to improve the quality of the Level 2 capacity estimates using additional or improved data and assumptions. However, the quality of the estimates does depend on the quality of the data on aircraft fleet mix, runway exit locations, and percentage of arrivals in the peak demand period. Time and Cost Requirements This method of capacity analysis requires only a small investment, commensurate with the fidelity of the capacity estimate obtained through this method. Completing a capacity analysis with this level of sophistication can be expected to take a few days to a week, and would likely cost less than $25,000. Model Availability Level 2 methods are publicly available as part of the AC, which is available for download from FAA’s website free of charge. A copy of the Prototype Airfield Capacity Spreadsheet Model devel- oped in this research and described in Chapter 4 of this guidebook is available on the CD-ROM provided with the guidebook. Level of Operator Skill and Training Required As with Level 1 models, users of these Level 2 charts/nomographs need to have a basic knowledge of airfield operations and the factors that affect airfield capacity, such as aircraft fleet mix, runway use configuration, and weather conditions. In addition to the Level 1 requirements, Level 2 users need to have an understanding of the specified percentages of arrivals, training flights (touch-and- goes), and the effects of the number and locations of exit taxiways on arrival runway occupancy times. With such a background, little or no training would be required to use the Level 2 charts and nomographs method. Summary of Limitations The main limitations of the currently available techniques at this level of sophistication are the limited flexibility inherent in the charts and nomographs and the unchangeable built-in assumptions associated with the 43 runway use configurations available. The values of hourly and annual capacity available in Chapter 3 of the AC may still reasonably apply to U.S. airports that have similar runway use configurations, but there is room for improvement and refine- ment at this level. Although this methodology presents greater flexibility than the table lookup method, in situations where more precise capacities are required—or if the conditions at the airport do not match the built-in default assumptions of the charts and nomographs—a higher level of modeling sophistication must be used. Level 3—Analytical Capacity Models This level of modeling sophistication refers to capacity (i.e., maximum throughput) analy- ses generated through analytical computer models, such as FAA’s ACM and the LMI Runway Capacity Model. Figure 3-3 shows a typical ACM graphical user interface.

38 evaluating airfield Capacity A typical analytical airfield capacity model can accept input assumptions and information on the following factors affecting airfield capacity: • Runway configuration • Types of operations (arrivals, departures, or both) assigned to each runway • Aircraft mix on each runway • Aircraft performance characteristics (e.g., minimum separation requirements, final approach speed, and runway occupancy times for arrivals and departures) • ATC rules and procedures (e.g., actual or standard separations, simultaneous runway occu- pancy), length of final approach, and flight rules based on ceiling and visibility (e.g., depen- dencies between runways and number of simultaneous movements) • Runway occupancy times and actual achievable average separation values, modeled as random variables (i.e., the user can specify a mean and standard deviation for these parameters from which buffers are estimated) The output of this model is an estimate of the hourly capacity of the runway system for any specified arrival-departure ratio or percentage of arrivals. A capacity arrival-departure envelope, which is often referred to as a Pareto frontier, also can be generated. Applications These models are applied in specialized airfield capacity studies, airport master plans, and regional airport system plans. This approach is limited to analysis of (1) the capacity of runways only (although taxiways and airspace constraints can be reflected implicitly), (2) systems of Source: Graphical User Interface (GUI) developed by Barrer Aviation Software. Figure 3-3. ACM graphical user interface.

existing airfield Capacity evaluation tools 39 runway configurations with moderate complexity and straightforward arrival and departure runway use procedures. Modeling Assumptions To use this model, the runway configuration must be available in the model, or be able to be represented by externally combining available configurations. Any limitations on aircraft types that can use a certain runway (e.g., because of length or noise restrictions) must be reflected externally to the model. A major assumption of this class of models is that taxiways and gates have little effect on determining airfield capacity. “Continuous demand for service” assumptions are associated with the configuration and operation of the departure and arrival airspace. Data Requirements The data requirements for analytical capacity models are more demanding than for the Level 1 or 2 models, yet not as complicated as those for a simulation-based model. Most data needed for Level 3 models can be compiled from readily available sources. Inputs include fleet mix, arrival and departure runway occupancy times, arrival-arrival separations, departure-departure separa- tions, and arrival-departure separations (i.e., how far out from the threshold an arrival must be to release a departure on the same or a dependent runway). Default values can be used for most of these inputs, excluding fleet mix, although the results would not be as precise as using data based on actual runway occupancy or separation through analysis of airborne and surface radar track data. In Level 3 modeling, a minimum amount of data will yield an answer, but this answer can become more refined or precise with the addition of more (and more airport- specific) data. Time and Cost Requirements This method of capacity analysis requires a moderate investment. Inputs to the models, such as the distribution for runway occupancy times and the standard minimum or actual aircraft separations, can be obtained from standard sources or can be developed through the processing and analysis of airborne flight path data (radar flight tracks) and surface track data (e.g., radar ASDE-X). Completing a capacity analysis using these analytical models can be expected to take a few weeks and would likely cost up to about $50,000. Model Availability The primary Level 3 model, ACM, is available free of charge from FAA. The other models in this category—the LMI Runway Capacity Model and FTA’s RUNCAP model—are proprietary. Level of Operator Skill and Training Required As with Level 1 and Level 2 models, users of Level 3 analytical capacity models need to have a basic knowledge of airfield operations and the factors that affect airfield capacity, such as air- craft fleet mix, runway use configuration, and weather conditions, as well as an understanding of the percentage of arrivals, training flights, and the effects of the number and locations of exit taxiways on arrival runway occupancy times. In addition to the Level 1 and 2 requirements, Level 3 users need to have an understanding of ATC rules and procedures and aircraft performance. Unlike Levels 1 and 2, a substantial amount of training is required to properly use the Level 3 analytical capacity models, including preparing model inputs, interpreting and analyzing model

40 evaluating airfield Capacity outputs, and representing airport-specific conditions implicitly. In addition, the Level 3 user must be more computer literate, because, at least in the case of the ACM, the model is not very user-friendly. Summary of Limitations This approach is limited to capacity analysis of the runways only (although taxiways and airspace constraints can be reflected implicitly), and is limited to analysis of systems of runway configurations with moderate complexity. FAA’s ACM also has the following limitations (which may or may not apply to other models in this category): • Some of the basic assumptions regarding arrival-departure separation requirements are hard- wired into the ACM legacy code. Any deviations from those assumptions require changes to and recompiling of that code. • Inability to directly account for taxiway and airspace constraints. These effects must be accounted for externally to the model and may require additional training. • Lack of flexibility in modeling runway configurations explicitly. Complex airfields must be analyzed by piecing together capacity estimates for components of the overall airfield opera- tion, and such post-analysis may require additional training. • Lack of flexibility in representing airfields that have runways with aircraft fleet mix restric- tions. A substantial amount of post-analysis is required to reflect such aircraft fleet mix restric- tions, which may require additional training. Level 4—Airfield Capacity Simulation Models This level of modeling sophistication refers to a level between the Level 3 analytical models, as represented by the ACM, and the Level 5 aircraft delay simulation models, as represented by SIMMOD, TAAM, and other models. This level of modeling sophistication allows the user to analyze complex runway use configurations with airport-specific ATC procedures and potential physical and environmental constraints. Two models that fit into this level are: (1) the MITRE runwaySimulator, and (2) the FTA Flexible Airport Simulation (FLAPS) model. Applications This level of modeling is best used to estimate an hourly throughput capacity of a runway system for a complex airfield, or for airport-specific operating procedures. New technologies and flight procedures can be represented in both models through use of the input parameters for aircraft operations, statistical buffers, runway dependencies, aircraft flight paths and profiles, and assumptions regarding required spacing between parallel runways. Both the runwaySimu- lator and FLAPS can include the effects of runway exit/entrance taxiways on airfield capacity. However, other taxiways on the airfield, such as parallel taxiways, connector taxiways, or runway- crossing taxiways, are not explicitly included in the Level 4 models. Modeling Assumptions To calculate a throughput capacity, Level 4 models do not use a detailed flight schedule. Instead, a saturated-conditions schedule is assumed, represented by a continuous arrival and departure stream in proportion to the fleet mix (which the user inputs). The arrival and departure stream is characterized by there always being aircraft waiting to land and take off (i.e., a continuous demand for service with no slack periods). The primary output of inter- est of the Level 4 models is runway throughput, not aircraft delay, although it appears that

existing airfield Capacity evaluation tools 41 the models can provide both. A typical output chart from the runwaySimulator is shown in Figure 3-4. Each plotted point represents the capacity estimate for 1 hour of the simulation at a certain arrival percentage. The clusters of plotted points are averaged to a centroid representing a particular percentage of arrivals, and these centroids are connected to draw a Pareto frontier or capacity curve. Until recently, runwaySimulator 2010 had two modes for modeling arrival runway occupancy times (AROTs). The first mode used an underlying trajectory landing-roll model, which incor- porated aircraft performance parameters for touchdown speed, deceleration, and exit speed. The second mode involved drawing an AROT from a user-specified distribution and recomputing the deceleration parameter that would realize it. Both modes had arrivals exiting at modeled exit locations. The version of runwaySimulator tested in ACRP Project 03-17 used a single mode, for which only the drawn AROT applied and the exit locations were ignored. The newest version of runwaySimulator, expected to be released in 2012, will support the two modes again, with the following adjustments: (1) use of the underlying trajectory landing-roll model will use exits, and (2) drawn AROTs will not. Data Requirements The Level 4 models require more detailed input data and assumptions than the Level 3 models, but less than the Level 5 models. The FLAPS model requires more detailed input data than does runwaySimulator because FLAPS explicitly models the aircraft landing, rolling out, and runway exiting process. The input stream contains essentially the same data as the Level 3 models, but at a greater level of detail. For example, close-in arrival and departure flight track geometries can be specified, and the user can specify aircraft fleet mix by runway. In Level 4 modeling, an answer can be obtained using a minimum amount of data, but the answer can become more refined or precise with the introduction of additional data. Source: MITRE Corporation. 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 Departures A rr iv al s Figure 3-4. Capacity curve or Pareto Frontier generated by the MITRE runwaySimulator showing clusters of points for each arrival-departure ratio.

42 evaluating airfield Capacity Time and Cost Requirements The investment required to apply a Level 4 model is somewhere between the investment required for the Level 3 analytical models and the investment required for the Level 5 aircraft delay simula- tion models, depending on the complexity of the problem. Because of their computational efficiency and simple input structures, the runwaySimulator and FLAPS models are characterized by a very low cost per computer run compared with Level 5 models. Typically, a Level 4 model takes several weeks to a month to set up and run, and an application of a Level 4 model would likely cost between $50,000 and $100,000, depending on the level of complexity of the issues being analyzed. Model Availability Currently, no model in this category is publicly available, as both runwaySimulator and FLAPS are proprietary. According to the MITRE Corporation, runwaySimulator will be made available to the public for a nominal licensing charge sometime in 2012. Level of Operator Skill and Training Required As with Levels 1, 2, and 3, Level 4 airfield capacity simulation models require operators to be familiar with airfield operations and the factors that affect airfield capacity, such as aircraft fleet mix, runway use configuration, and weather conditions. Operators must also have an under- standing of the percentage of arrivals, training flights, the effects of the number and locations of exit taxiways on arrival runway occupancy times, and an even more detailed understanding of ATC rules and procedures and aircraft performance in landing and takeoff. Compared with the Level 3 models, proper use of the Level 4 models requires extensive training that includes preparing model inputs, interpreting and analyzing model outputs, calibrating the model, and representing complex runway operations and dependencies between movements. Taxiway Configurations Both the runwaySimulator and the FLAPS model consider only the runways and exit/entrance taxiways, not the other taxiways on the airfield, such as parallel taxiways, connector taxiways, or runway-crossing taxiways. However, parallel taxiways can be modeled implicitly through the input parameters for runway occupancy times, and crossing taxiways can be modeled implicitly through other input assumptions. Variability The runwaySimulator has six sources of randomness: flight generation, arrival runway occu- pancy times, departure runway occupancy times, arrival release times, departure release times, and times between departure release and start of roll. The developers of the runwaySimulator experimented with random buffer sizes at one point but abandoned that approach. Instead, sta- tistical variability of aircraft separations is represented implicitly in runwaySimulator by apply- ing the assumed statistical excess-spacing buffers to the minimum separation requirements. FLAPS is a stochastic, event-driven simulation model that produces statistical outputs on runway capacity and use, aircraft delays, exit use, and runway queues. Other Factors The runwaySimulator software runs on a personal computer (PC) running Microsoft Win- dows® and currently requires runtime licenses from Wolverine Software for the SLX and Proof

existing airfield Capacity evaluation tools 43 Animation (Proof) software used by the current simulation engine (i.e., runwaySimulator 2010). However, MITRE expects to eliminate the requirement for a separate SLX license in the version of the runwaySimulator that is expected to be made public in 2012. Summary of Limitations The main limitations of the Level 4 models are that they consider runways only and exit/ entrance taxiways, not the entire taxiway system, except that both models can reflect, at least implicitly, (1) runway exit taxiways and parallel taxiways through the definition of runway occupancy times, and (2) entrance taxiways through potentially longer-than-typical departure- departure separations. Level 4 models also require a more sophisticated user with a greater understanding of complex airfield operations and simulation models. Level 5—Aircraft Delay Simulation Models Aircraft delay simulation models represent the highest level of modeling sophistication for evaluating runway capacity and measuring aircraft delay within a single modeling envi- ronment. Historically, simulation models were developed to analyze complex airport and airspace operating environments where multiple factors, such as airfield configuration, ter- minal aprons, airspace limitations, and aircraft activity, interact in ways that simpler models cannot represent. Multiple examples of simulation models have a long history of use over the past 30 years. Most recently, TAAM and SIMMOD have been the most widely used models (see Figure 3-5). However, RDSIM, ADSIM, and the AirTOp Fast Time Simulator by AirTOpsoft are also in cur- rent use. The following characteristics differentiate simulation models from less sophisticated models: • They represent aircraft activity as a flight schedule showing aircraft travel through the airport (e.g., landing, gate-in, gate occupancy, gate-out, and takeoff). • They have the ability to model apron-gate operations and aircraft taxiway movements. • They use networks to represent the airport’s surface configurations (i.e., runways, taxiways, and gates) and networks or flight routes to represent airspace configurations (i.e., approach and departure routes). The ability of Level 5 simulations to model detailed aircraft movements along taxiways and through the airspace makes them more difficult and time-consuming to set up, run, and calibrate than the lower-level models. Some Level 5 models provide user flexibility to scale down the level of fidelity for a particular application (e.g., users can operate with runways only or with gate areas instead of individual gates). Applications Level 5 models are best used to analyze complex capacity issues involving aspects of airfield operations not focused on the runways (i.e., aprons, taxiways, and airspace), or to analyze inter- actions between multiple aspects of the airfield. Simulation models provide a high degree of fidelity for capacity issues that may be subject to intense public scrutiny. Level 5 models also are needed if estimates of metrics such as aircraft taxi time and delay are required. In situations where visual validation and computer animation graphics are needed, the use of Level 5 models would also be required. Finally, Level 5 models are required for the analysis of detailed flight schedules.

44 evaluating airfield Capacity Modeling Assumptions Very few built-in assumptions are associated with Level 5 models; most inputs are variables, so that users can define their specific airfield and airspace situation through the model inputs. Data Requirements Simulation models require a high level of detail about aircraft operations and airfield and airspace configurations, which requires an extensive data gathering and analysis effort. As with lower levels of modeling, in Level 5 modeling a minimum amount of data will yield an answer, but this answer can become more refined or precise with the introduction of additional data. Even more than at previous levels of modeling, however, the outputs of a Level 5 model will only be as good as the input data and assumptions provided by the user. Time and Cost Requirements Simulation models take significant resources to set up and calibrate. Data collection and model calibration take a significant amount of time. Detailed discussions with air traffic and air- port operations specialists and airport users are often required to properly represent the baseline Source: LeighFisher. Figure 3-5. Screen shot generated by TAAM showing traffic to and from Dallas/Fort Worth International Airport.

existing airfield Capacity evaluation tools 45 operating conditions and proposed changes. However, once a simulation model has been set up and calibrated, it can be run for additional cases fairly quickly. Thus, the most cost-effective use of simulation modeling is when multiple problems are anticipated to be evaluated so that the cal- ibrated model can be used more than once. Level 5 simulation models may take several months to be properly set up and calibrated to provide representative capacity and delay estimates, and they would likely cost more than $100,000 to apply to a complex airport and set of experiments. Model Availability Most simulation models are not available free of charge to the public. One exception is the base version of FAA’s SIMMOD, which is available from FAA free of charge; other versions of SIMMOD must be purchased or leased from the model vendor, ATAC Corporation. Similarly, TAAM licenses must be obtained from Jeppesen/Boeing on a monthly basis. Level of Operator Skill and Training Required As with Level 4 models, users of Level 5 aircraft delay simulation models need to be very familiar with airfield operations, factors that affect airfield capacity, the effects of traffic and run- way configurations on occupancy times, ATC rules and procedures, and aircraft performance in landing and takeoff. In addition to the Level 4 requirements, Level 5 users need to have a detailed understanding of aircraft performance on the taxiway system and in the apron-gate area, and detailed knowledge of airline flight schedules. Compared with the Level 4 models, the proper application of Level 5 models requires even more extensive training, including preparing model inputs, interpreting and analyzing model outputs, and representing complex runway, taxiway, and apron-gate operations, in a greater level of detail than for Level 4. Taxiway Configurations Level 5 aircraft delay simulation models are the only tools available to model taxiway and other airfield operations that do not directly support runway operations, which could include parallel taxiways, circulation taxiways, bypass taxiways, crossover taxiways, apron-edge taxiways, apron taxilanes, and holding bays. Variability Most simulation models incorporate a number of random variables to account for the natural variability in flight schedules, aircraft performance, and airport operations. Most simulation models do not incorporate random variability in aircraft taxiing speeds and tend to model all aircraft taxiing operations at the same speed, although different taxiway-speed groups can typi- cally be defined for different taxiway categories or locations. Some Level 5 simulation models simplify aircraft landing roll performance modeling, limiting the ability to model this key airfield capacity determinant. Other Factors Complex aircraft delay simulation models should be calibrated against actual data on aircraft taxiing times, aircraft flow rates, and aircraft delays. These comparative data generally can be obtained from FAA’s Aviation System Performance Metrics (ASPM) database; however, the user must be careful to obtain the aircraft delay data that are most comparable to aircraft delay esti- mates produced by the aircraft delay simulation models, which basically can be defined as excess travel times or the times that aircraft spend waiting to land, waiting for a gate, waiting to push

46 evaluating airfield Capacity back, waiting to depart, and so forth. Calibration is critical to ensure that the logic in the model, most often related to runway loadings and aircraft separation, is not more efficient or precise than what human controllers and pilots could realistically achieve. Summary of Limitations The main limitations of simulation models are the time and cost required to set up and cali- brate the models and the requirement for a well-trained and knowledgeable user. These models also involve a large number of variables, and there is a considerable learning curve to use the models and check the results.