Need for Data: A User s Perspective SESAR WP-E TREE project Carlos Regidor, May 13 th EUROCONTROL ART WS 01/15 Validation/Measuring ATM Performance
OBJECTIVES Development of a simulation model capable of: Recreating and predicting the delays propagation in the Network and; Analysing delay mitigation strategies applied by either the Airlines or the Network Manager. SCOPE Characterization of the propagation of Reactionary Delays taking into account the influence of the aircraft links and passengers and crew connections; Implementation and testing of specific Airlines Disruption Management Strategies; Evaluation of the Network and Airlines planning Performance; 1
WHY DO WE MODEL? The ATM system contains a huge amount of heterogeneity and correlations; The only sensible technique to generate realistic simulations and testable predictions is to use a data-driven approach. This means that data collected on the conditions of the ATM system and its different components is employed during the simulation process; The need of data is thus an absolute priority in this type of modelling. 2
WHY DO WE NEED THE DATA? Holistic use of data: Performance of the simulations Validation and calibration of the model predictions Simulation outputs should be comparable to the performance data contained in the reference data sources 3
WHAT DATA DO WE NEED? Flight Performance Data: Daily flight plans with scheduled/actual arrival and departure times of all the flights along with the aircraft rotations CODA Market Sector Data: Passenger connections between the different flights arriving and departing from each airport SABRE Primary Delays: Declared primary delays whenever they affect the first flight of an aircraft rotation CODA Airport Nominal capacities: Eurocontrol Public Airport Corner + DDR2: from input traffic samples or public sources 4
DATA NEEDS FOR MODEL VALIDATION and CALIBRATION The Flight Performance Data is required not only for traffic simulations but for the validation of the model predictions; The model is able to predict if a large congestion would be developed given the daily schedule and the primary delays. Each model realization will generate a pattern of delay propagation from the primary delays on; It will include the list of affected flights, the airports they visited in their rotations as well as the status of each airport in terms of congestion along the day hour by hour; Clusters of congested airports The simulation outputs should be comparable to the performance data contained in CODA; 5
WHAT PROBLEMS DID WE FIND? Flight Performance Data: Cumbersome and time consuming process Access granted only to small traffic samples and with several limitations Miss the potential capacity of the model to process bigger data Other sources: Flightstats.com ; Flightradar24.com, etc Reliable enough? Market Sector Data: Data not available for the general audience Expensive data for R&D purposes General: Sharing of data between WP-E projects proved to be not so effective 6
Isdefe C/ Beatriz de Bobadilla, 3 28040 Madrid Tel.: +34 91 411 50 11 Email: general@isdefe.es www.isdefe.es