The Modernized NASPAC Simulation Environment

1
The Modernized NASPAC Simulation Environment
2
Your Fathers NASPAC
National Airspace System Performance Analysis
Capability

• The FAAs standard system-wide model
• Originally developed in late 1980s
• Represents NAS as network of interconnected
  queues
• Airports
• TRACONs
• En route sectors
• Discrete-event
• SIMSCRIPT II.5
• Fortran, C, Pascal pre-processors
• Sun Solaris platform
• Used mainly for investment analysis (i.e.,
  cost-benefit analysis)

3
Legacy NASPAC Environment
TAF Terminal Area Forecast OAG Official
Airline Guide ETMS Enhanced Traffic Management
System
4
Objectives of Modernized NASPAC

• Make available to more analysts and projects
• Maintain fast run time
• Incorporate traditional trajectory module with
  fuel burn computation
• Automate airport weather scheduler
• Obtain modern, configuration-controlled,
  documented source code
• Produce standard, analyst-friendly, repeatable
  output
• Validate
• Develop GUI and output visualization tools
• Facilitate Monte Carlo simulation

5
Modernized NASPAC Environment
Wx
Delay
Fuel Burn Rate
TAF
Future Schedule Generator
ETMS
Constrained Traj.-Based Forecast (2D)
4D Trajectories
Unconstrained Traj.-Based Forecast (2D)
Legacy NASPAC
Sector Entry/Exit Times
Flight Table
Industry/ Regulatory Response
Itinerary Generation
Fleet Evolution
Java Trajectory Module
Sector Crossings
Output Parser
Core Queuing Model
Airport Table
Sector Table
Unsatisfied Demand
Wind Field
Sector Geometries
Fleet Forecast
Flight Delays
Sector Capacities
Pareto Curves
Capacity Scheduler
Wx
Wx weather TAF Terminal Area Forecast MAP
Monitor Alert Parameter ETMS Enhanced Traffic
Management System
Airport Capacities
MAP Values
Flow Restrictions
6
Future Traffic Generation
TAF

• FAA airport forecasts are unconstrained
• Yield excessive and unrealistic future delays at
  some airports
• Estimated airport capacities used to constrain
  traffic growth
• Mimic operator and FAA response to large demand
  and capacity imbalances
• Apply both schedule smoothing and trimming at 110
  airports

7
Itinerary Generation

• Constrained schedule consists of ETMS flight legs
• No tail numbers
• Create aircraft itineraries
• Link flight legs with same airline and aircraft
  type
• Minimize total fleet size, i.e. number of
  itineraries
• Respect schedule turn time constraints

ATL
Flight Leg arriving ATL
A
Flight Leg Departing ATL
B
Flight Leg Departing ATL
C
Min schedule turn time
8
Fleet Evolution

• Retire and replace aircraft with newer types
• minimize changes across seat-class/engine-type
  categories
• Consistent with APO and Mitre Fleet Forecast
• Domestic passenger and cargo airlines only

9
Java Trajectory Module

• Aircraft performance
• BADA 3.6 tables
• Waypoints / cruise alt.
• ETMS flight plan
• Arrival/departure fixes
• Assigned to flight path
• Wind data
• NCEP/NCARGlobal Reanalysis Model
• 4D track points computed at 1 min. intervals

10
Airspace Modeled

• 939 airspace elements

11
Airports Modeled

• All IFR flights touching U.S. airspace are
  modeled
• Arrival/departure capacities for 110 airports
  included
• All other airports assumed to have infinite
  capacity
• VFR flights at 73 airports included

12
Output Files
Input Files
Perl Scripts
Airport Capacity File (airportc) ATO-P Input
Schedule pre-configuration File Sector Capacity
File (sector_sim) sim-configuration File
Flight File
Oracle Database
Output Files
Airport File
Arrival Trace File(t_arr) Arrivals-Departures at
Sink Trace File(t_snk) At-Gate Trace
File(t_agt) Change Parameter File
(change_apt) Cross Arrival Fix-Departure
Fix-Restriction Trace File(t_cnd) Departure Trace
File(t_dep) Enter-Exit Sector Trace
File(t_sec) IFR Trajectories(traject) JTM Bad
Flights File JTM Flight Map File JTM Fuel Usage
File JTM Generate Itineraries Schedule Pushback
Trace File(t_pbk)
Sector File
13
Metrics

• Flights accommodated
• Flight trimming process
• Delay
• Gate push-back
• Departure runway/fix queuing
• Sector queuing
• Arrival fix/runway queuing
• Fuel Burn
• Origin to Destination
• US airspace
• CO2
• Jet A savings converted to CO2 using multiplier
  of 21.095 lb/gal

14
NASPAC Project Wiki
15
Validation Process
Every new software release validated!

• Run simulation for eight historical days
• Demand variation
• Two days per season (1 weekend, 1 weekday)
• Weather variation
• Actual weather used for each day
• Distribution of surface and en route weather
• Compute flight time, delay, fuel burn metrics
• Compare validation metrics to observed system
  response
• Ground Truth ASPM, ASQP, BTS databases
• Mean, variance
• Pooled, pair-wise statistics

16
Airborne Delay, Daily Observations
v5d
17
Airborne Delay, OEP Airports
ORD
FLL
v5d
19 Oct. 2006
18
Taxi-Out Delay, Daily Observations
v5d
19
Push-Back Delay
v5d
20
ETE Distributions, 19 Oct. 2006
v5d
21
Airport Arrival Rate, 19 Oct. 2006
Charlotte-Douglas International Airport
v5d
22
Fuel Burn
23
Conclusions

• Modernized model runs much faster, accessible to
  more users, and easier to use
• Deterministic trajectory model provides easy
  means of specifying flight trajectories
• Uses standard Eurocontrol BADA data
• Fuel burn computation integrated into trajectory
  model
• Accommodates wind field
• Achieves good accuracy in nominal flight time and
  fuel burn
• Airborne delays close to observed values
• Surface delays appear underestimated

24
Next Steps

• Evolve International carrier fleet
• Enhance airport capacity model
• Incorporate additional airport pareto curves
  based on alternative configurations and/or wind
  conditions
• Improve taxi-out delay
• Update departure fix en trail constraints
• Incorporate Traffic Flow Management (TFM)
• Ground Delay Programs
• Include Monte Carlo simulation capability
• Validation, validation, validation

25
Questions?