Macroscopic Workload Model for Estimating En Route Sector Capacity

Transcription

1 Paper No. 37 Macroscopic Workload Model for Estimating En Route Sector Capacity July 4, 2007 J. D. Welch, J.W. Andrews, B. D. Martin B. Sridhar NASA Ames Research Center Workload Capacity Paper 37 1

2 Introduction Airspace capacity estimates are key flow management tools Hot topic: complexity and dynamic capacity reductions How about static, un-reduced capacity? The starting point for all dynamic density estimates Should be easily determined But we still use subjective estimates to establish max capacity Controllers declare parameters based on operational experience Unfortunately, controllers don t really know best! Low traffic demand guarantees that many large en route sectors never operate at maximum capacity High local traffic densities limit many small sectors to operate well below their optimistic subjective capacity thresholds Fortunately, an existing simple workload model can predict the maximum capacity of any en route sector Its predictions match peak US sector counts exactly Workload Capacity Paper 37 2

3 Model Overview Uses Schmidt s* simple sector workload equation Gives workload as a function of traffic density We choose parameters to minimize complexity No vertical rates Longitudinal flow We set workload to maximum safe level We invert equation to determine corresponding max traffic density Sector capacity depends on flow direction and sector geometry Analytical model simplifies fitting to observed peak sector traffic Bounds maximum sector traffic counts * Schmidt, David K., A Queuing Analysis of the Air Traffic Controller s Work Load, IEEE Transactions on Systems, Man, and Cybernetics, Vol. SMC-8, No. 6, June Workload Capacity Paper 37 3

4 Macroscopic Workload Model G = G b + G c + G r + G t Sector Workload Intensity Background Conflict Recurring Transition Gray: original model* Black: new elements** as conflicts arise periodic at sector crossings 1 Traffic limit for sector Traffic density G c = τ c [B κ (κq + 1)] G r = τ r [κq/p] G t = τ t [κq/t] service times (measured ) Sector Volume occurrence rates (calculated from airspace parameters) Workload Intensity Model Parameters G b = 0.1 τ c = 50 s τ r = 2 s τ t = 15 s P = 300 s T = 1200 s M h = 7 nm M v = 1000 ft E[V 21 ] = 440 kt Q = 10,000 nm 3 Background Gb Human Workload Limit G = Gb+Gt+Gr+Gc Conflict Gc Transition Gt Recurring Gr Workload Capacity Paper Aircraft Count * Schmidt, David K., A Queuing Analysis of the Air Traffic Controller s Work Load, IEEE Transactions on Systems, Man, and Cybernetics, Vol. SMC-8, No. 6, June ** Andrews and Welch, "Workload Implications of Free Flight Concepts", 1st USA/EUROCONTROL ATM Seminar, Saclay, France 1997.

5 Determining Service Time Parameters Traditional approach Observe controller performance times via real time simulation Schmidt* (τ c and τ t ), Majumdar** (τ t ) Macroscopic approach Invert workload equation to calculate capacity bound N max Calculate N max for a large number of sectors Fit service times to match N max to observed peak sector traffic * Schmidt, David K., A Queuing Analysis of the Air Traffic Controller s Work Load, IEEE Transactions on Systems, Man, and Cybernetics, Vol. SMC-8, No. 6, June ** Majumdar, Arnab and Ploak, John, Estimating Capacity of Europe s Airspace Using a Simulation Model of Air Traffic Controller Workload, Transportatoin Research Record 1744, Paper No , Workload Capacity Paper 37 5

6 Sector Capacities from Model Sector Traffic Count (aircraft) ZDC32 Q = 10,684 nm 3 H = 9,900 ft L ~ 170 nm Nm = 16.3 aircraft Model Capacity (All (455 CIWS CWIS* Sectors) ZTL45 Q = 10,297 nm 3 H = 13,800 ft L ~ 80 nm Nm = 12.9 aircraft 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 Sector Volume (nm 3 ) Model Parameters G b = 0.1 τ c = 50 s τ r = 2 s τ t = 15 s P = 300 s T = L/550 kt M h = 7 nm M v = 1000 ft E[V 21 ] = 440 kt Workload Capacity Paper 37 6 Capacity varies as ~ (sector volume) 1/2 At a given volume, elongated sectors have higher capacity * Corridor Integrated Weather System Domain in Northeast US

7 FAA Monitor Alert Parameters - CIWS Sectors 25 ETMS Monitor Alert Thresholds vs Sector Volume - CWIS Sectors 20 MA Threshold (aircraft) Sector Volume (nm 3 ) Monitor Alert Parameter ~ independent of sector volume Workload Capacity Paper 37 7

8 CIWS Sector Peak Traffic Aircraft Count Aircraft Density Super High High Low Workload Capacity Paper 37 8

9 Model Capacities and Observed Counts (Original Parameter Estimates) Sector Traffic (aircraft count) All CIWS Sectors Model capacity (peak count) Peak clear-weather traffic count 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 Sector Volume (nm 3 ) Original parameter estimates tt= 15 transit work time/aircraft (s) tc= 50 conflict work time/aircraft (s) tr= 10 recurring work time/aircraft (s) Mh= 5 (miles) Workload Capacity Paper 37 9

10 Fitted Model Capacities and Observed Counts Sector Traffic (aircraft count) All CIWS Sectors Model capacity (peak count) Peak clear-weather traffic count 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 Sector Volume (nm 3 ) Fitted Model parameters Maximum observed count at each volume fits model bound. Most sectors operate below capacity Workload Capacity Paper tt= tc= tr= Mh= 15 transit work time/aircraft (s) 50 conflict work time/aircraft (s) 10 2 recurring work time/aircraft (s) 5 7 (miles)

11 Observed Counts vs Model Capacities All CIWS Sectors Peak Clear-Weather Traffic Count Model Capacity Few sectors operate above the model bound Workload Capacity Paper 37 11

12 Traffic Densities CWIS Sectors smaller than 10,000 nm 3 Density (aircraft/nm 3 ) Model capacity (peak density) Observed peak clear-weather density ,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000 Sector Volume (nm 3 ) Small sectors handle denser traffic (but with more controllers per aircraft) Workload Capacity Paper 37 12

13 Model Workload Components All CIWS Sectors Conflict workload Gc Transition workload Gt Recurring workload Gr Workload Intensity ,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 Sector Volume (nm 3 ) At capacity, conflict workload dominates in smaller sectors Workload Capacity Paper 37 13

14 Observations from Model Most en route sectors operate below model capacity lack of demand little-used routes system flow constraints local deviations from model parameters higher airspeeds (closing, transit) non-longitudinal flow increased separation standards (horizontal, vertical) Large sectors improve workplace efficiency (aircraft/controller) Small sectors handle higher traffic densities Workload Capacity Paper 37 14

15 Conclusions Macroscopic workload model fits observed peak traffic data operationally reasonable service times all sector shapes and volumes Explains observed relationships utility of small sectors in dense airspace non-linear sector capacity growth with volume Applications replacement for declared capacity real-time sector capacity estimation tool for airspace design guidance for ATM research Workload Capacity Paper 37 15

16 Next Steps Include traffic information Vertical motion Will increase conflict frequency Mean transit time Will increase transit frequency (usually) Examine more sectors U.S. Europe Workload Capacity Paper 37 16

17 END Workload Capacity Paper 37 17