Use of DDR data in PREDICT to support the pre-tactical planning (D-6 to D-1) FMP Exchange Workshop, edition th of September 2012

Transcription

1 Use of DDR data in PREDICT to support the pre-tactical planning (D-6 to D-1) FMP Exchange Workshop, edition th of September 2012

2 Presentation content: DDR concept overview. Predict evolutions using DDR data - NM 16.X Release Impact on predictability, issues & lessons learned: OPS evaluation results (March to May 2012).EURO 2012 (4 coordinated AD in Poland / in Ukraine) Olympics 40 coordinated airports in the London area Next improvements 2

3 Multi-components concept Francis airline schedules from Innovata 98% of airlines worldwide weekly main sources of flight intentions airport slot data from EUACA approximately 180 coordinated airports Daily DDR/PRISME : collect and process flight intentions from multiple sources (incl. IATA ICAO conversion) Storage of historical traffic samples Long term forecasts generated by users seasonal forecasts generated by users Long term & Seasonal forecasts DDR1 extensions providing: traffic forecasts for long term & seasonal PREDICT pre-tactical forecasts access to NEST (SAAM + NEVAC) pre-tactical forecasts Pre-tactical planning D+1 to D+6 forecasts enriched with flight intentions & NAT flow substitution PREDICT Pre-tactical forecasts pre-tactical forecasts from PREDICT enhanced interface internet SITA line ANSP able to download: user generated long term forecasts user generated season forecasts pre-tactical reference forecasts from PREDICT historical traffic samples NEST Terminology Long term typically refers to interval a few years before operations to one year before operations user access to DDR2 Web Portal CHMI FMPs able to visualise / download flight lists & entry counts from PREDICT relating to pre-tactical forecasts that exploit DDR2 flight intentions from EUACA 3

4 Predict evolutions - NM 16.X Release Flight Intentions Data: multi-sources DDR 2 Schedules Management process Collected Intentions Airport slots Consolidate flight intentions Matching IATA / ICAO flight Id s Forecast Traffic samples, enriched with intentions D-6 to D-1 Other D-6 to D-1 PREDICT D+1 to D+6 Model traffic & airspace demand Enrichment of historical data Route assignment NAT flows substitution (automated) at D-3 Network collaborative pre tactical planning Francis Historical Flight Data RPL Data D-6 updates Airspace structure description DDR portal DDR Users (FMP s) 4

5 Post-op comparison: classic versus new PREDICT counts compared to actual counts Number of flights in the network. Day to day number of flights. Where are the flights, are they in the correct part of the network? Number of flights per ACC. Hourly sector counts. 70+ sectors targeted for detail analysis. Aerodromes, coordinated v uncoordinated. Aerodromes, top 50 plus interested NMC ones. 4D route assessment. Manual check on added flights. 5

6 Operational assessment results Hourly counts predictability variation (06-20:00 UTC): New Predict versus Classical Predict Predictability variation= % Predictability New Predict - % Predictability Classical Predict Week following the clock change Easter Week Easter WE +7days Bank Holiday WE s Bank Holiday WE s Bank Holiday WE s 6

7 COUNTS PER HOUR IN ROLLING HOURS EVERY 20 MIN Francis Operational assessment results Illustration 1 COMPUTE_TIME ID EDDF 120 Actual New (DDR) Classic PREDICT Week of the clock change and start of summer season 0 * NO PATCH APPLIED 7 00:00 00:20 00:40 01:00 01:20 01:40 02:00 02:20 02:40 03:00 03:20 03:40 04:00 04:20 04:40 05:00 05:20 05:40 06:00 06:20 06:40 07:00 07:20 07:40 08:00 08:20 08:40 09:00 09:20 09:40 10:00 10:20 10:40 11:00 11:20 11:40 12:00 12:20 12:40 13:00 13:20 13:40 14:00 14:20 14:40 15:00 15:20 15:40 16:00 16:20 16:40 17:00 17:20 17:40 18:00 18:20 18:40 19:00 19:20 19:40 20:00 20:20 20:40 21:00 21:20 21:40 22:00 22:20 22:40 23:00 23:20 23:40 Data Sum of TFC_A Sum of TFC_C Sum of TFC_N START_TIME

8 COUNTS PER HOUR IN ROLLING HOURS EVERY 20 MIN Francis Operational assessment results Illustration 2 COMPUTE_TIME ID EDDF 120 Actual 100 New (DDR) Classic PREDICT 20 0 Echoes of the strike in EDDF ground operations on the 27/03/12 * NO PATCH APPLIED 8 00:00 00:20 00:40 01:00 01:20 01:40 02:00 02:20 02:40 03:00 03:20 03:40 04:00 04:20 04:40 05:00 05:20 05:40 06:00 06:20 06:40 07:00 07:20 07:40 08:00 08:20 08:40 09:00 09:20 09:40 10:00 10:20 10:40 11:00 11:20 11:40 12:00 12:20 12:40 13:00 13:20 13:40 14:00 14:20 14:40 15:00 15:20 15:40 16:00 16:20 16:40 17:00 17:20 17:40 18:00 18:20 18:40 19:00 19:20 19:40 20:00 20:20 20:40 21:00 21:20 21:40 22:00 22:20 22:40 23:00 23:20 23:40 Data Sum of TFC_A Sum of TFC_C Sum of TFC_N START_TIME

9 PREDICT enrichment of Historical flight data using Airports slot data collected in DDR2 OPS deployment from 04/06 Purposes: Support Euro in June, in Poland and the Ukraine. Olympics (21/07 to 15/08) 9

10 Euro 2012 DAILY MOVEMENTS: UKCC DONETS K AIRPORT UKCC A/D (All) NEW (DDR) ESP v POR Actual FRA v ENG UKR v FRA ENG v UKR ESP v FRA D+7, 23/6 100 Classic PREDICT 50 0 Predictability deviation: DDR enrichment versus Classic Predict 07/6 08/6 09/6 10/6 11/6 12/6 13/6 14/6 15/6 16/6 17/6 18/6 19/6 20/6 21/6 22/6 23/6 24/6 25/6 26/6 27/6 28/6 29/6 30/6 01/7 SOURCE 1.Intentions_D 4.Historical_D 5.Historical_D-7 10

11 Olympics 2012 Overall performance of pre-tactical forecasts, all airports in London area Actual, this year DDR forecasts Actual, last year Deviation: Forecasts - Actual 11

12 Issues during the Olympics: Issue 1 ACL updates not communicated to EUACA (1 week before 21/07) FI sources change punctual duplication of FI UID corrective measure applied removing from predict obsolete FI Issue 2 Duplicates (same ARCID for different flights, close timing) Issue 3 Pseudo Duplicates (same ARCID for different flights, different timing) Issue 4 Flight intentions planned but not operated or inaccurate 12

13 Issue 2: Duplicates (same ARCID for different flights, close timing = deviation <= 30 min) 1 to 3 % duplicates 13

14 Issue 3: Pseudo Duplicates (same ARCID for different flights, different timing = deviation > 30 min) Pseudo-duplicates = inaccurate result of the IATA to ICAO Flight ID matching process same (i-match) arcid, timing deviation > 30 min different IATA flight id (only visible in PREDICT HMI Flights list, not in CHMI Flights list) 14

15 Issue 4: Flight intentions planned but not operated or inaccurate Forecast counts Actual counts (post-op) Flight intentions planned but not operated 15

16 Next improvements? Corrective measures of detected issues (lessons learned) Data Steward function systematic process supervision Enrichment process using additional sources : - INNOVATA schedules data - Greek Airport data - Demand data about SKI axis (Winter season) - AD slots from Turkey - AD Demand Data for special events at non coordinated airports (e.g. Cannes / Monaco events) Route assignment improvements - improved city pairs stat (AO) 16

17 Yours views? For later questions or newly detected anomalies in PREDICT data Please mail it to: 17

18 Reserve slides 18

19 PREDICT Enrichment process: Francis -Collect flight intentions, Integrate updates -IATA / ICAO conversion, -Transform repetitive intentions into daily FI ICAO / IATA flight ID s bridge data Schedules Airport Slots Other Flight Intention data DDR2 FI data management Matching IATA / ICAO flight Id s Consolidated Flight Intentions (C.F.I.) automated D-6 to D-2 + manually at D-1 if required Planned flight intentions not correlated with historical data Added Historical Flights not correlated with the F.I., from trusted airports Historical unconfirmed flights, but from not trusted airports + unconfirmed RPL s Enriched historical traffic sample Historical Traffic Sample (TS) PREDICT Enrichment of historical data with DDR data Potential Deletes Unconfirmed Confirmed Added Unconfirmed Confirmed Route assignment Forecast Traffic samples enriched with C.F.I. New RPL s Deletion process 19

20 PREDICT route assignment Added Route assignment Unconfirmed Confirmed Forecast Traffic samples enriched with C.F.I. Priority order: historical route from similar flight from the same AO ( similar: same airport pair ) City Pairs Stat: most frequently flown route during the current/last Airac cycle Route generated by NM Path Finder Direct - Keep the historical route (FTFM) - RPL: last available / updated route - FI values supersede historical values (optionally activated) for ADEP, ADES, EOBT, AC type 20

21 PREDICT Deletion process Added Potential Deletes Unconfirmed Deletion process Added Unconfirmed Trusted airports (or airlines)? Depending the estimated coverage of the collected intentions, Estimated coverage is computed on demand Confirmed Confirmed Trusted airports: NMC experts judgment Expert judgments are defined per airport ( or per airlines, in future), Purpose: restrict the deletion process to - trusted airports (or airlines, in future) and for all flight types - Possibility to exclude some flight types, namely: General aviation, Others", Military flights 21

22 Euro 2012 DAILY MOVEMENTS: EPGD -- GDANSK Francis AIRPORT EPGD A/D (All) ESP v 350 ITA NEW (DDR) GER v GRE Actual ESP v IRL CRO v ESP Classic PREDICT Predictability deviation: DDR enrichment versus Classic Predict 0 07/6 08/6 09/6 10/6 11/6 12/6 13/6 14/6 15/6 16/6 17/6 18/6 19/6 20/6 21/6 22/6 23/6 24/6 25/6 26/6 27/6 28/6 29/6 30/6 01/7 SOURCE 1.Intentions_D 4.Historical_D 5.Historical_D-7 22