Presentation is loading. Please wait.

Presentation is loading. Please wait.

USING HISTORICAL FLIGHT DATA TO EVALUATE AIRBORNE DEMAND, DELAY AND TRAFFIC FLOW CONTROL Michael Brennan, Terence Thompson Metron Aviation, Inc Steve Bradford,

Similar presentations


Presentation on theme: "USING HISTORICAL FLIGHT DATA TO EVALUATE AIRBORNE DEMAND, DELAY AND TRAFFIC FLOW CONTROL Michael Brennan, Terence Thompson Metron Aviation, Inc Steve Bradford,"— Presentation transcript:

1 USING HISTORICAL FLIGHT DATA TO EVALUATE AIRBORNE DEMAND, DELAY AND TRAFFIC FLOW CONTROL Michael Brennan, Terence Thompson Metron Aviation, Inc Steve Bradford, Diana Liang FAA/ASD 100

2 2 Overview Basic problem and objectives Methodologies Flight delay estimation Weather related volume and delay Visualization of volume and delay propagation Terminal demand and en route delay Spatial and temporal propagation of delay Effects of ATC initiatives

3 3 Airborne Demand and Delay Improving Traffic Flow Management and Air Traffic Control procedures requires understanding where and when delay occurs and how it relates to demand Previous efforts have had to focus on single flight behavior or on delays inferred from extended time on route Needed to develop techniques to –Identify where and when flights take delay en route –Aggregate delays over space and time for visualization and analysis

4 4 Methodologies Determine time and place each flight takes delay en route Aggregate per-flight delay over space and time Spatial and temporal grid –Divide airspace into 15 min x 15 min spatial blocks –Divide time into 5 minute segments –Consider high and low altitudes (partition at 20k feet) –Map per-flight volume and delay into these cells Use this and other representations for visualization and analysis

5 5 Identifying Per-Flight Delay Initial analysis inferred delay from changes in ETA estimates in ETMS TZ messages –Upon receiving each TZ, ETMS evaluates the flight’s progress with respect to the flight plan –If progress is not as expected the ETA is recalculated and the new estimate is sent as part of the TZ –Changes in ETA for a flight should then indicate where and when delays are taken Approach was flawed –Too many random changes in ETMS ETA estimates –Too much ‘noise’ in the analysis

6 6 Reconstruction of En Route Delay Developed new techniques for determining flight-specific en route delay –Based on analysis of TZ position reports evaluated against currently active flight plan –Captures delays due to vectoring, circular holding and reroutes Supports evaluation of the relationships between volume, delay, and its propagation

7 7 En Route Delays are Computed by Comparing TZs to Flight Plan Proceeding to Waypoint No delay Vectoring.2 min delay / min of flight Perpendicular to flight path 1 min delay / min of flight Anti-parallel to flight path 2 min delay / min of flight Path of AC and TZ point Path of flight plan and way point

8 8 Results of New Delay Estimation Techniques Circles indicate where estimated en route delay increases Rectangles indicate where estimated en route delay decreases Delay calculations based on ETMS ETAs are subject to random fluctuation Delay calculations based on new techniques are much more stable

9 9 Volume Buildup Behind Weather System Volume of NY-Bound Traffic over ½ Hour Period Heavy Traffic from Mid-West Cannot Penetrate Weather

10 10 Localization of Weather Related Delays is Clearly Identified Delays for NY-Bound Traffic over ½ Hour Period Flights Delayed While Waiting for Weather to Pass Flights Routed Around Local Disturbances

11 11 Geographical Visualization of Airborne Volume High volume of low altitude ORD-bound flights at Southwest arrival fix at 2000Z … ORD

12 12 2015Z 2010Z 2000Z 2005Z Geographical Visualization of Delay Propagation … leads to propagation of high-altitude airborne delay back over 1000 miles by 2015Z ORD

13 13 Airborne Maneuvering to Slow Traffic 15 minute flight segments in 5 minute blocks Circles indicate delay recorded on posit Shading indicates delay in cell

14 14 Excess Arrival Demand Creates Airborne Delay in the Terminal Area Delta arrival banks Airborne delay in terminal area

15 15 Terminal Area Delay Back-Propagates Through the Airspace

16 16 Quantitative Evaluation of Delay Propagation Conjectured that pockets of high delay lead to further delay upstream at later times –Can this hypothesis be demonstrated? Approach –Identify high-delay cell at T 0 –Identify flights that will pass through cell after T 0 –Model distance by flight time to establish range rings –Compute average flight delay D i,j at each time period T i in each range ring R j

17 17 Initial high delay in cell at time T 0 Delay in ring R 1 at T 0 is D 1,0 Delay in inner ring at T 1 is D 0,1 Delay in ring R 1 at T 1 is D 1,1 Spatial and Temporal Modeling for Delay Propagation Identify flights bound for high delay cell

18 18 Time After Peak Delay (T i ) Flight Time Range Ring (R i ) 510152025303540450 5 10 15 20 25 30 35 0 Spatial and Temporal Propagation of Delay

19 19 Affect of ATC Delays on Peak Volume Cells 0 5 10 15 20 25 30 Volume in Cell Frequency of Occurrence Delayed Undelayed 0 1 2 3 4 5 6 7 8 9 Volume in Cell Frequency of Occurrence Delayed Undelayed All TrafficORD Flights ATC Delays reduce occurrence of highest volume cells

20 20 Next Steps These techniques are relatively new – many applications unexplored Identify recurring patterns of delay – suggest changes in –Aircraft routing –TFM and ATC procedures –Airline scheduling –Airspace design Identify instances and patterns of under and over control Add delay estimation to ETMS process –Notification to users of flight delays –Real-time display of delay maps in ASDs

21 21 End of Presentation


Download ppt "USING HISTORICAL FLIGHT DATA TO EVALUATE AIRBORNE DEMAND, DELAY AND TRAFFIC FLOW CONTROL Michael Brennan, Terence Thompson Metron Aviation, Inc Steve Bradford,"

Similar presentations


Ads by Google