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
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
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
Simulation Airspace and Traffic Model RAMS 5.04 Plus 5 States scenario (Dowdall 2001) 12 Sep. 1997 =100% 150% (~2005) 200% (~2010) 300% (?) 140 sectors from 24 ATC Karlsruhe, Maastricht and Reims 36 en-route sectors above flight level 245 9. R. Dowdall, 2001, 5 States Fast-Time Simulation, EUROCONTROL Experimental Centre, EEC Report No. 361, www.eurocontrol.int/eec/publications.html 4
Number of Aircraft per Hour for Three Measured Centres (150% Scenario) 900 800 700 600 500 400 300 200 100 0 00:00 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
Conflict-Attitudes Per Flightlevel Bands for the 3 Centres 300 250 200 150 100 50 0 190 210 230 250 270 290 310 330 350 370 390 410 Cruise-Cruise Cruise-Attitude Attitude-Attitude 6
12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% Aircraft at flight level of entire 1997 scenario 7 40-49 70-79 100-109 130-139 160-169 190-199 220-229 250-259 280-289 310-319 340-349 370-379 400-409 10-19
Performance Parameters Of Speed Control Rüdiger Ehrmanntraut Frank Jelinek EUROCONTROL Experimental Centre (EEC) Brétigny sur Orge, France 8
Average (cruise) speeds per FL with std.-dev. 600 500 400 300 200 100 0 9 10-19 30-39 50-59 70-79 90-99 110-119 130-139 150-159 170-179 190-199 210-219 230-239 250-259 270-279 290-299 310-319 330-339 350-359 370-379 390-399 410-419
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 % 0 200 220 240 260 280 300 320 Flight Level 340 360 380 400 999 3. [12] A. Nuic, 2004, BADA Version 3.5, EUROCONTROL Experimental Centre, www.eurocontrol.fr/projects/bada/ 10
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
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
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
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
Results- Overview 16000 100% 14000 12000 10000 90% 80% 70% 60% 8000 50% 6000 4000 2000 40% 30% 20% 10% 0 A B C D E A B C D E A B C D E A B C D E 100 150 200 300 Conflicts Unresolved % Resolved Unresolved / Flights 0% 15
Results- Distribution of Speed Adjustments # resolved 1200 1000 800 600 400 200 0-18 17-16 - 15-14 - 13-12 - 11-10 - - 9-8- 7-6- 5-4- 3-2- 1+ 1+ 2+ 3+ 4+ 5+ 6+ 7+ 8+ 9 10 + 11 + 12 + 13 + 14 + 15 + 16 + 17 + 100 - B 100 - D 100 - E 150 - B 150 - D 150 - E 200 - B 200 - D 200 - E 300 - B 300 - D 300 - E Speed Adjustment % many small speed adjustments + and unbalance is due to rule base, not to ac performance 16
Results- Encounter Angles and their Resolution Rates 8 000 100% 7 000 6 000 80% 5 000 60% 4 000 3 000 40% 2 000 1 000-100 150 200 300 100 150 200 300 100 150 200 300 100 150 200 300 B C D E 0-10 10-30 30-90 % resolved 0-10 % resolved 10-30 % resolved 30-90 20% 0% Other study shows that conflicts almost equally distributed over angle 1-179 More speed resolutions for wide angles Small angel = 0 and 180 Climbs and descents can still be resolved with speed 17 0 90 180
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
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
Results- Conflict Clusters Average Sizes 10000 Log count Normal conflict Resolution rates 90% 80% 1000 70% 60% 100 50% 40% 30% 10 20% 10% 1 100 150 200 300 100 150 200 300 100 150 200 300 100 150 200 300 100 150 200 300 100 150 200 300 100 150 200 300 100 150 200 300 100 150 200 300 100 150 200 300 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
Results- Number Constraining Aircraft (1) 900% PC vertical TC nil 700% 500% PC speed TC nil 300% 100% 0% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 100 200 300 # constraining Traffic growth vertical and speed resolution differ 21
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 9 300 Read: xx % of conflicts with n constraining aircraft could be resolved by TC-PC 22
Results- Environmental Parameters for TC speed PC speed 101% 100% 99% 98% 97% 96% 95% 94% D D D 100 150 200 Fuel CO HC NOx H2O CO2 SO2 small effect or even positive effect of speed control 23
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 What is Lateral Offset?
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
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
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
Results (1) Rejoin Rules Setup 1: offset-distance 5 + n*2 <15 NM, offset angle 30 deg +/- 15 Resolution trigger time -15 minutes 6000 5000 4000 3000 2000 1000 80% 70% 60% 50% 40% 30% 20% 10% Year 2010 0 Y-combi Y-EoC Y-EoR Y-EoS Y-TOD 200 0% Conflicts Unresolved % Resolved Unresolved / Flights 29
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 /178-180 2-15 /165-178 15-165 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
Results (3) Different Offset Anges 68.0% 66.0% 64.0% 62.0% 60.0% 58.0% 56.0% 10 15 20 25 30 35 40 45 Scenrios run each with different offset angle Offset angle = rejoin angle Observed optimum at 35 degrees 31
Results (4) Discrete Offset Distance 6000 5000 4000 3000 2000 1000 0 1 x 7NM 2 x 7NM 3 x 7NM 5+n*2<15 200 80% 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
Results (5) Discrete Offset Distance Distribution 10000 1000 log scale 100 10 1-21 -14-7 0 7 14 21 0 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
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 Results (7) Picture Simulation Output
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
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 BAW414 290 AFR8910 LL 290 330 AF747 330 ABC DEF BAW4561 330 S B S A BAW5944 330 GHI DHL1234 335 370 OPR Cab ac cd AZA1234 335 370 SAB910 ST 289 260 LTU114 NE 335 370 AFR745 310 LMN S XYZ RST UVW 37
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
Lateral Offset or Speed? 16000 90% 14000 12000 10000 8000 6000 4000 2000 80% 70% 60% 50% 40% 30% 20% 10% 0 LO Speed LO Speed Traffic 200 = 2010 Traffic 300 = 2025 0% Conflicts Unresolved Resolutions % Unresolved/Flights % Attention: two models are compared Both perform very well LO better than speed control? 39
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