Modeling and Simulation
Challenges and the New
Vehicle NRA
Ken Wright
Sensis Corporation
January 28, 2010
Background
 Analyze the combined impacts of advanced vehicles and
NextGen
2
Vehicles
SST
CESTOL
VLJ
Cruise @ M 0.78,
35k – 40k ft
Cruise @ M 0.6,
35k ft
Cruise @ M 0.5,
25k ft
UAS
LCTR
Cruise @
M 0.28,
15k ft
400 nm
750 nm
1200 nm
2000 nm
4000 nm
3
Cruise @ M 1.6
41k ft to 53k ft
Background (Continued)
 Timeframe
– 2025
– 2040 (roughly 2 times today’s traffic levels)
– 3X (3 times today’s traffic levels)
 Tasks
– Design vehicles
– Develop usage scenarios
– Compute delay
– Compute environmental impacts
– Compute metroplex impacts
– Compute safety impacts
– And much more
4
Model Infrastructure
NextGen CONOPS,
SAE 1845 AIR procedures (<10kft),
BADA Performance (>=10kft) conv. AC
FACET
Sensis New Vehicle
NRA Model
Interaction Diagram
BADA
BADA
Arrival, departure
procedures for new
vehicles
Procedures
JPDO-provided
NextGen
configuration
GT
Safety
Models
SIMMOD
ACES
Flyability
Analysis
BADA
Procedures
Demand
TARGETS
BADA
Procedure
Design
Procedures
New Vehicle
Aircraft Design
Impacts
Procedures
Demand
Procedures
BADA
JPDO-provided
NextGen
configuration
Demand
Demand files
Air Taxi
Demand File
TSAM
BADA
AvDemand
Travel Population
Costs
Growth
BADA AC Performance, New Vehicles
EDS
BADA AC Performance
BADA AC Performance, New Vehicles
BADA
Flight Demand
Growth
Procedures
Vehicle Market Penetration,
Fleet Mix Changes
4D Trajectories Regional
4D Trajectories System-wide
Safety
Metrics
Noise
Contours
AEDT
System
Wide
PDARS
Safety
Metrics
Emissions
Inventories
NAS-Wide Airport and
Airspace Performance
with and without New
Vehicles for NextGen
APMT
5
Emission
Impacts
AEDT
Regional
Noise
Impacts
Regional
Airport and
Airspace
Performance
New Vehicle
Regional Airport
and Airspace
Design Impacts
Safety
Impacts
Key:
Text in black Input or output to a model
Text in blueResult from a model
The Focus of this Presentation is ACES
• To do NAS-wide simulation in ACES we need:
• BADA Data
• Schedules/Capacities
• ACES tracks used by AEDT to compute environmental
impact
6
General Modeling Simulation Observations
 Designing good experiments is tough
– Results often seem obvious
– Difficult to translate concepts into hypotheses that leverage model
strengths
 Difficult to spot erroneous results
– Large datasets, complex interactive models
– Figuring what’s going on below the surface is difficult
 Some modules don’t work well with large datasets
– Results not significantly impacted
– New features can increase run time, without affecting results
– ACES 5.0 vs ACES 6.2 compatibility issues
7
Challenge: Demand Generation
JPDO-Provided Demand Sets
9
Huge delays at LAS and ATL due to
capacity-demand imbalance dominate
systemwide-results
12
ACES-Specific Observations
 MPAS (physics model) currently can’t fly unconventional
vehicles
– Cruise-Efficient STOL Aircraft (cruises too slow)
– Supersonic transport (cruises too fast)
– Tiltrotor (takes off like a helicopter)
 Tail-tracking
 Terminal area tracks
 Conflict Detection and Resolution
 Weather obstacles (Back-up)
13
Tail-Tracking
 One-third of all delay is propagated delay
 ACES tail-tracking doesn’t quite work
14
Sensis Itinerary-Creation Algorithm
gives 5-6 Stages per Airframe
15
Increase in delay due to tail-tracking
too small to be realistic
16
Aircraft is in two places
at the same time
1
2
3
4
Short flights show discontinuous tracks
 Flights having great circle distance less than about 60 nm fly
discontinuous “X” routes
 Affects 3,000 to 3,500 ACES flights in 2040 dataset (about 3.5% of
total)
 Because flights fly longer routes than necessary, amount of time
spent airborne is too long.
18
19
20
21
Flight from TEB to LGA is
Airborne for 31 Minutes.
Average speed = 21 mph
22
Conflict Detection and Resolution
 Not used in New Vehicle project
 Causes about as many conflicts as it resolves
23
CDR Causes Collision
CDR Causes Collision
26
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
Aircraft sense impending 5-nm conflict
CDR Causes Collision
Aircraft move to avoid 5-nm conflict
CDR Causes Collision
Aircraft move to avoid 5-nm conflict
CDR Causes Collision
Boom!
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
CDR Causes Collision
3X-Dataset
 No restrictions on international arrivals leads to large sectors
exceeding capacity limits
57
2086
Name
ZNY87
ZBW01
ZHU79
ZNY90
Capacity
25
35
25
25
Foreign arrivals choke off
domestic departures.
(175k flights)
Conclusions
 Big project; tested lots of models; pushed limits of ACES and AEDT
 It’s probably more fun to be a developer than a user
 Coming up with interesting experiments is tough
– Experiments should leverage model strengths to find novel results
 Difficult to spot erroneous results
 Some modules/capabilities did not work as expected for our large-
scale experimental configuration
60
Questions?
Ken Wright
Sensis Corporation
Back up
Rerouting around Weather
 Mostly worked, but number of weather polygons scenario could use
was unnecessarily limited
 Peculiar reroutes due to lack of airborne holding capability in ACES
63
Clusters of Storm Cells
(>=50 DBZ)
Replaced by Rectangles
64
Clusters of Storm Cells
(>=50 DBZ)
Replaced by Rectangles
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
Final fix blocked
80
Final fix blocked
81
Final fix blocked
82
83
84
85
86
87
88
89
90
91
It appears that the flight turns
around only when it bounces
off another weather cell
92
93
94
95
96
97
98
99
10