4th EUROCONTROL Innovative Research workshop Day II December 2005, the 7th Performance Parameters Of Speed Control & Potential Of Lateral Offset

Transcription

1 4th EUROCONTROL Innovative Research workshop Day II December 2005, the 7th Performance Parameters Of Speed Control & Potential Of Lateral Offset Rüdiger Ehrmanntraut EUROCONTROL Experimental Centre (EEC) Brétigny sur Orge, France 1

2 Strategic Traffic Organisation Concept Elements Strategic Traffic Organisation CE1 Complexity Prediction CE2 Multi Sector Traffic Organisation CE3 Functional Airspace Segregation CE4 Synchronous Planning Conflict Geometries Lateral Offset Highways Contracts Conflict Density Vertical Offset 4D / 3D Tubes System Wide Information Management Flow and Sequences Speed Control Dynamic Sectors Cluster Organisation 3D Airways 2

3 Validation process Concept definition Concept element 1: Complexity Concept element 2: MSP Concept element 3: 3D routes Fundamental ATM research Why this approach? What is it? What benefit? Capacity, safety, environment Fast time simulation System Design How? Automation Information system Technologies 3

4 Simulation Airspace and Traffic Model RAMS 5.04 Plus 5 States scenario (Dowdall 2001) 12 Sep =100% 150% (~2005) 200% (~2010) 300% (?) 140 sectors from 24 ATC Karlsruhe, Maastricht and Reims 36 en-route sectors above flight level R. Dowdall, 2001, 5 States Fast-Time Simulation, EUROCONTROL Experimental Centre, EEC Report No. 361, 4

5 Number of Aircraft per Hour for Three Measured Centres (150% Scenario) :00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00 KARLSRUHE MAASTRICHT REIMS 5

6 Conflict-Attitudes Per Flightlevel Bands for the 3 Centres Cruise-Cruise Cruise-Attitude Attitude-Attitude 6

7 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% Aircraft at flight level of entire 1997 scenario

8 Performance Parameters Of Speed Control Rüdiger Ehrmanntraut Frank Jelinek EUROCONTROL Experimental Centre (EEC) Brétigny sur Orge, France 8

9 Average (cruise) speeds per FL with std.-dev

10 Aircraft Performance Envelopes Update with BADA 3.5 (Nuic 2004) More than 100 performance models Precise speed models for 20 aircraft types Other speed profiles by ICAO class: Heavy, Medium, Light Limited to 15% from nominal behaviour 30% 25% 20% 15% 10% 5% 0% A30B Cruise Speed Variance in % Flight Level [12] A. Nuic, 2004, BADA Version 3.5, EUROCONTROL Experimental Centre, 10

11 Conflict and Controller Model ac2 e2 e4 ac1 e1 e3 S2 S1 Planner (PC) LPC = 15 min SPC = 7NM Tactical (TC) LTC = 0 min STC = 5NM Separation Buffer S Look-ahead parameter L Maximal Implementation interval R 11

12 Rulebase Logic Conflict Geometry Analysis Which a/c to penalise? Speed Reduction? Speed Increase? Set Resolution Constraints New speed, times Penalise: 1. requesting occupied stable FL 2. nearer to airport 3. inbound and second 4. inbound vs outbound 5. evolution vs stable 6. behind 7. below 8. in descent 9. significantly faster 10. about to leave cruise 11. still on ground Reduce penalised Increase penalised Reduce favourite Increase favourite Set resolution start time Set resolution stop time Iterate speed 12

13 Uncertainty Model Uncertainty modeled by higher lateral separations Planner separated on 7NM horizontal buffer, 1000 feet vertical Planner 6% uncertainty per hour = 1.5% in the look-ahead horizon. Radar 0% uncertainty 7 NM 5 NM 13

14 Simulation Scenarios A Conflict Detection only, no resolutions B C D E Planner resolves speed-only Radar is disabled Planner resolves vertical-only (+/- 2 RVSM flight levels) Radar is disabled Planner resolves speed-only Radar resolves speed-only Planner resolves speed-only Radar resolves horizontal-vectors-only 14

15 Results- Overview % % 80% 70% 60% % % 30% 20% 10% 0 A B C D E A B C D E A B C D E A B C D E Conflicts Unresolved % Resolved Unresolved / Flights 0% 15

16 Results- Distribution of Speed Adjustments # resolved B D E B D E B D E B D E Speed Adjustment % many small speed adjustments + and unbalance is due to rule base, not to ac performance 16

17 Results- Encounter Angles and their Resolution Rates % % % % B C D E % resolved 0-10 % resolved % resolved % 0% Other study shows that conflicts almost equally distributed over angle More speed resolutions for wide angles Small angel = 0 and 180 Climbs and descents can still be resolved with speed

18 Results- Minimal Displacement Distance and their Resolution Rates 90% Resolution rates 80% 70% 60% 50% 40% 1NM 2NM 3NM 4NM 5NM 6NM 7NM B C D E CPA Distance critical conflicts more difficult safets? The PC-speed-only and the PC-speed-TC-vectors 18

19 Results- Conflict Clusters Cluster: transitive closure of conflicting aircraft in time (and distance). Cluster A-B-C If D in conflict with A, or B, or C, then cluster A-B-C-D A B C D 19

20 Results- Conflict Clusters Average Sizes Log count Normal conflict Resolution rates 90% 80% % 60% % 40% 30% 10 20% 10% CL_1 CL_2 CL_3 CL_4 CL_5 CL_6 CL_7 CL_8 CL_9 CL_10 0% Cluster size Senario D = planner and radar use speed-only Cluster = +/- 5 FL and 8 minutes, no distance limitation Exponential decrease of cluster sizes Graceful degrading for big clusters 20

21 Results- Number Constraining Aircraft (1) 900% PC vertical TC nil 700% 500% PC speed TC nil 300% 100% 0% # constraining Traffic growth vertical and speed resolution differ 21

22 Results- Constraining Aircraft Average Resolution Rates 100% 80% TC speed PC nil TC speed PC vector 60% 40% TC speed PC speed 20% TC vertical PC nil 0% r 0 r 1 r 2 r 3 r 4 r 5 r 6 r 7 r 8 r Read: xx % of conflicts with n constraining aircraft could be resolved by TC-PC 22

23 Results- Environmental Parameters for TC speed PC speed 101% 100% 99% 98% 97% 96% 95% 94% D D D Fuel CO HC NOx H2O CO2 SO2 small effect or even positive effect of speed control 23

24 Discussion + Conlcusions Capacity: Speed control very high performance Graceful behaviour for conflict clusters Graceful behaviour for conflict with many constraining aircraft Automation with MSP If MSP function is automated and uses speed control only, then the radar controller is left with much less conflicts, but with complicated clusters. Automation: The capacity barrier is the limited airspace - efficient packing and traffic organisation are required Safety Speed resolutions have good behaviour for critical conflicts Safety best when planner and radar controller apply different resolutions (orthogonal solutions) Radar control did not impact the ability of a long-term planner (15 minutes = MSP) to resolved with speed control Environment Small speed adaptation often sufficient No or very little positive impact when applying speed control 24

25 25 What is Lateral Offset?

26 Lateral Offset - Literature Herndon, A. A., et al., 2003 Utilizing RNAV Avionics: Testing Lateral Offset Procedures, MP03W160, The MITRE Corporation, McLean, VA., in proceedings of the 22nd DASC, Indianapolis, Indiana, USA field trials in Albuquerque and Houston, single centre: 32 aircraft Controllers and pilots seem to accept procedure Herndon, A. A., J. S. DeArmon, J. Spelman, 2004 Use Of Lateral/Parallel FMS Procedures And Implementation Issues, MITRE CAASD, in proceedings of the 23rd DASC, Salt Lake City, USA Minneapolis ARTCC: 15 flights Quantification of benefits difficult due to small traffic set Reconfirms acceptance Related literature: RNP-RNAV, ICAO SASP 26

27 RAMS Simulation Model AT (time) TURN RIGHT/LEFT (heading or number of degrees) OFFSET (distance) REJOIN ROUTE ABEAM (location) Start time, time increment, search direction Offset direction: left, right Offset angel Offset distance Rejoin location: end of sector, end of centre, end of route, top of descent tcd T t LO implementation tcs T t Tlook ahead 27

28 Lateral Offset Resolution Model 1 2 T 3 4 FOR EACH rejoin-location (rule-base driven) FOR flight1 AND flight2 (rule-base driven) FOR forward-search OR backward-search FOR EACH offset-distance FOR EACH offset-direction FOR EACH offset-angle TRY new trajectory 28

29 Results (1) Rejoin Rules Setup 1: offset-distance 5 + n*2 <15 NM, offset angle 30 deg +/- 15 Resolution trigger time -15 minutes % 70% 60% 50% 40% 30% 20% 10% Year Y-combi Y-EoC Y-EoR Y-EoS Y-TOD 200 0% Conflicts Unresolved % Resolved Unresolved / Flights 29

30 Results (2) Resolution Rates per Encounter Angel and Geometries 100% 90% 80% 70% 60% 50% Climb- Climb Cruise- Cruise Descent- Descent Climb- Cruise Climb- Descent Cruise- Descent 0-2 / / all best is descent-descent worst is climb-climb cruise is insensitive for encounter angle Cruise-Desc 16% Climb- Descent 8% Climb-Cruise 33% Climb-Climb 6% Cruise- Cruise 33% Desc-Desc 4% 30

31 Results (3) Different Offset Anges 68.0% 66.0% 64.0% 62.0% 60.0% 58.0% 56.0% Scenrios run each with different offset angle Offset angle = rejoin angle Observed optimum at 35 degrees 31

32 Results (4) Discrete Offset Distance x 7NM 2 x 7NM 3 x 7NM 5+n*2< % 70% 60% 50% 40% 30% 20% 10% 0% Conflicts Unresolved Resolutions % Unresolved/Flights % Resolution with 1, 2, or 3 offset distances in 7 NM intervals 2*7 and 3*7 equal 1*7 worst, but still very high resolution 32

33 Results (5) Discrete Offset Distance Distribution log scale deviation 1 deviation 2 deviations 3 deviations 4 deviations Resolution with 1offset distances in 7 NM distributes until +/- 21NM No aircraft encountered >4 deviations A 10% rule? Resolution space still drastically under-utilised 33

34 Results (6) Uncertainties 80% 70% 60% 50% 40% PC 4% PC 6% PC 12% PC 18% TC 4% TC 12% TC 18% % Resolutions Conflict detection with 4, 6, 12, or 18% uncertainty Lateral Offset requires long look-ahead times Planning Controller PC degrades Radar Controller TC degrades faster 34

35 35 Results (7) Picture Simulation Output

36 Found Issues Syntax needs improvement proposed clearance can lead to error Corrected in this study Syntax needs improvement - clarity on offset distance When aircraft already flying LO get another additional LO clearance We propose always absolute distance from initial flight plan LO direction should be optimised for trajectory 36

37 Discussion (1) Multi Sector Planner as Conflict Solver: LO very high performance LO insensitive for uncertainties LO low impact on flight efficiency Complexity Reduction LO has very high degree-of-order (metric proposed by Ehrmanntraut) Flow Safety Maximisation LO safer because conflict densities reduced LO safer because of planning horizon, time to chaos is bigger BAW AFR8910 LL AF ABC DEF BAW S B S A BAW GHI DHL OPR Cab ac cd AZA SAB910 ST LTU114 NE AFR LMN S XYZ RST UVW 37

38 Lateral Offset Conclusions LO very high resolution rates optimum offset angle 35 degrees best for cruising en-route traffic operates well with reduced number of offset distances operates well under high uncertainties there are open issues very useful for a tactical Multi Sector Planner should improve safety is easy and cheap to implement 38

39 Lateral Offset or Speed? % % 70% 60% 50% 40% 30% 20% 10% 0 LO Speed LO Speed Traffic 200 = 2010 Traffic 300 = % Conflicts Unresolved Resolutions % Unresolved/Flights % Attention: two models are compared Both perform very well LO better than speed control? 39

40 Summed Up Both measures very promising Sector capacity booster To discuss Quantify workload and capacity? Operational feasibility to let MSP operate with speed and offset? Uncertainties due to look ahead horizon? 40

41 41