Phase 2 report - Project Final Report

Transcription

1 MINT Phase 2 report - Project Final Report Project Number: SJU/LC/0021-CTR Project Title: Minimum CO 2 in TMA Work package: 4 Date: Status: Draft Classification: Internal Author(s): Christer Forsberg, AVTECH

2 Document Identification Programme: Project Number: Project Title: Project Acronym: Project Manager: Partners: SJU / Aire SJU/LC/0021-CTR Minimum CO 2 in TMA MINT MINT Mr. Christer Forsberg, AVTECH Christer.forsberg@avtech.aero Consortium partners: AVTECH Sweden AB LFV Group Novair Airbus Egis Avia Expert Advisory Group partners: GE Aviation Thales Avionics Thales Air Systems SAS Document title: Project Final Report Document Id: AVTECH_MINT_WP4_Project Report Work Package No: 4 Classification: Internal Date: Christer.forsberg@avtech.aero Organisation maintaining document: AVTECH Executive Summary This document contains the project final report for the SJU/AIRE sponsored project MINT and is the deliverable for the second and last phase of the project. During 2009 the MINT project (Minimum CO 2 in Terminal Manoeuvring Area) consortium partners demonstrated optimized aircraft operation during descent with 11 Novair A321 flights into Stockholm Arlanda airport. The optimization was made by addressing both the lateral as well as the vertical parts of the approach. For the lateral part an RNP-AR procedure was developed for Arlanda Runway 01R which minimize track miles while considering noise sensitive areas. The vertical profile was optimized by uplink of individual flight selected descent wind information to enable the aircraft Flight Management System (FMS) to select the best possible Top of Descent (ToD) point in order to achieve an idle continuous descent approach. Last but not least, five of the flights were flown with a time constraint to a point near FL100 in order to demonstrate the aircraft capabilities to fly an efficient descent while also supporting a sequencing time request from ATC as foreseen in the SESAR operational concept. To be able to collect results from different flights arriving during different periods and with different environmental conditions the aircraft collected data was normalized for the differences in cruise altitude, aircraft mass, descent winds and ISA deviation. After the normalisation of the MINT flights as well as for a series of earlier collected aircraft logs, representing the baseline scenario, it was identified that the results Classification: Internal Page 2 of 74

3 in terms of fuel savings was 165 kg when comparing flights arriving only from the south into Runway 01R. This corresponds to more than 500 kg of CO 2 savings per approach. The normalisation process however also enables all flights to be compared even though arriving from different directions and into different runways. When comparing the 11 MINT flights to a series of 28 Novair flights into Arlanda airport the average saving for the MINT flights resulted in 140 kg of fuel savings per approach. The project also analyzed the noise distribution based on the aircraft collected data and the results show clearly that the RNP-AR procedure has big possibilities to address noise by circumnavigating sensitive areas. For Arlanda airport, noise problems currently limits the operation during periods of the day into Runway 01R and further research with the input from the new procedure will be made with the ambition to remove these limitations. The project have analyzed the saving potential for not having to circumnavigate the airport to use Runway 19L/R instead of being able to use Runway 01R when approaching from the south and found that in the order of 300 kg of fuel per approach can be saved corresponding to almost 1 metric ton of CO 2. 1 Operationally there is no problem identified with implementing the new procedure and it is actually planned to go into normal operation during January Initially it will, however, only be possible to use it during low traffic periods due to mixed equipage issues during the other periods. The initial 10 flights were carried out during the summer 2009 on normal revenue flights. After data collection and initial analysis was carried out, a dissemination event was arranged where the invited people was allowed to participate on the last demonstration flight that was a dedicated demonstration flight. The dissemination event was very well attended and overall the project has received a high interest and received well coverage in media. The MINT project have demonstrated minimum CO 2 in the terminal manoeuvring area through optimized arrivals by combining benefits from using the aircraft RNP capability with benefits from flying efficient continuous descent approaches. It has also demonstrated the possibility to use these techniques to address noise issues as well as showing that it is possible to combine them with the SESAR defined CTA concept. Distribution List Name MINT Focal point contacts Alain Siebert Michael Standard Partners SESAR JU SESAR JU Organisation 1 Full combustion of 1 kg of fuel corresponds to 3.15 kg of CO 2. Classification: Internal Page 3 of 74

4 Control Page This version supersedes all previous versions of this document. Version Date Author(s) Pages Reason Christer Forsberg All New document Christer Forsberg All Input from Tom Maier, Flavien Gueudre, Lionel Bernard-Peyre, Patrick Manzi and Henrik Ekstrand. Completely updated version that qualifies for final draft Christer Forsberg All Final input from partners as well as Alain Siebert from SJU. Added executive summary Christer Forsberg Mostly editorial updates Classification: Internal Page 4 of 74

5 Table of Content CONTROL PAGE INTRODUCTION SCOPE OF THE DOCUMENT TEAM COMPOSITION REVISIONS PROJECT OBJECTIVES HIGH-LEVEL OBJECTIVES DEMONSTRATION OBJECTIVES QUANTIFY THE POTENTIAL OBJECTIVES DISSEMINATION OBJECTIVES TRIAL PREPARATIONS AND EXECUTION FLIGHT TRIAL PREPARATIONS Getting the authority s acceptance Aircraft preparations Airline preparation Development of the Operational and Data Collection procedure Special preparations for the Dissemination event flight FLIGHT DEMONSTRATION SCENARIO Set-up Execution Data collection Performing the trials ANALYSIS Efficiency analysis Wind uplink accuracy analysis Navigation accuracy RTA accuracy Noise analysis Collection of future ground system functionality requirements RESULTS EFFICIENCY NAVIGATION Lateral analysis Vertical profile WIND UPLINK RTA ACCURACY NOISE DISTRIBUTION EFFECTS COLLECTION OF FUTURE GROUND SYSTEM FUNCTIONALITY REQUIREMENTS RESULTS AND OBSERVATIONS FROM ATC POSSIBLE DEPLOYMENT SCENARIO AND RELATED RESULTS SUMMARY OF DISSEMINATION ACTIVITIES CONCLUSIONS, RECOMMENDATIONS AND FINAL WORDS ANNEX A: PROCEDURE FOR ARLANDA RWY 01R ANNEX B. AVTECH NORMALIZATION PROCESS DESCRIPTION PROVIDED FOR MINT Classification: Internal Page 5 of 74

6 ANNEX C. DISSEMINATION EVENT INVITATION AND AGENDA ANNEX D. PRESS RELEASES ANNEX E. SAFETY CASE MATERIAL DOKUMENTINFORMATION: ANNEX F. LINKS TO MINT PROJECT MENTIONED IN MEDIA Table of Figures Figure 1. All 10 MINT flights as well as all 28 baseline flights Figure 2: MINT Flights RNP Approaches Figure 3: Lateral Deviation below feet Figure 4: Lateral Deviation from SASA695 (FAF) Figure 5. Comparison between the 10 MINT flights and a sample of the reference flights Figure 6. FMS prediction results based on inputs from AVENTUS and flight planning system compared to actual winds Figure 7: MINT Waypoint location Figure 8. Noise impact for the suburb Upplands Väsby as well as the hospital Lövenströmska Figure 9: Noise footprint comparing ILS and RNP-AR procedures into Runway 01R Figure 10. Inhabited areas compared to ILS and RNP-AR procedures into Runway 01R Figure 11. Noise distribution compared for demonstration flight and a simulated traditional flight path Figure 12. Radar screen showing the MINT flight in green on approach to Runway 01R Figure 13. Radar plot showing first MINT flight and another flight (NDC506) Figure 14. Airports by size and number of known runways Figure 15. RNAV (RNP-AR) procedure for Arlanda airport Runway 01R Figure 16: Direct route identification Figure 17 Vertical profile example Figure 18 Descent Identification Figure 19 Descent profile examples Figure 20 Wind impact on the flown distance during dt seconds Figure 21 Wind triangle Figure 22: Actual and normalized fuel burn results during different wind conditions Figure 23 Temperature effect on altitude Figure 24 Trajectories with different ISA deviations Figure 25: Temperature ISA deviation during Novair A321 approach Figure 26: Novair RNP flight - ISA normalization made with BADA process Figure 27: Different ToD for different mass Figure 28: TAS of flights with different mass Figure 29: Mass normalization for the 45 tons flight Figure 30: Mass normalization for the 50 tons flight Figure 31: Mass normalization for the 59 tons flight Figure 32: Mass normalization for the 63 tons flight Figure 33: Flights flying on different altitude Figure 34: Altitude normalization Simulation flying on FL Figure 35: Altitude normalization Simulation flying on FL Table of Tables Table 1. Analysis of MINT flights compared to other baseline flights into other runways Table 2. Result of the FMS wind prediction correspondence to actual wind profiles based on inputs provided by AVENTUS and traditional flight planning system respectively Table 3. Various possible services levels depending on air and ground equipments Classification: Internal Page 6 of 74

7 List of Abbreviations ACARS ACMS AIRE AMAN ANSP AOC ATM ATC ATCC CDA CDM CIES CMU CO 2 CTA DAR EC EIS ETA FAF FIR FL FMS IAF ICAO ILS IRP ISA KPI NM QAR R&D RNAV RNP RNP-AR RTA R/T RWY SJU STAR TMA ToD TP TS VOR WP 4DT Aircraft Communications Addressing and Reporting System Aircraft Condition Monitoring System Atlantic Interoperability Initiative to Reduce Emissions Arrival Manager Air Navigation Service Provider Airline Operational Centre Air Traffic Management Air Traffic Control Air Traffic Control Centre Continuous Descent Approach Collaborative Decision Making Collaborative Information Exchange System Communication Management Unit Carbon dioxide Controlled Time of Arrival Digital ACMS Recorder European Commission Electronic Instrument System Estimated Time of Arrival Final Approach Fix Flight Information Region Flight Level (Altitude [feet]/100) Flight Management System Initial Approach Fix International Civil Aviation Organization Instrument Landing System Initial Reference Point International Standard Atmosphere Key Performance Index Nautical Mile (1852 meters) Quick Access Recorder Research and Development Area Navigation Required Navigation Performance Required Navigation Performance Authorisation Required Required Time of Arrival Radio Transmission Runway SESAR Joint Undertaking Standard Terminal Arrival Route Terminal Manoeuvring Area Top of Descent Trajectory Predictor the Swedish Transport Agency Very High Frequency Omni directional Range (ground based navigation aid) Work Package 4 Dimensional Trajectory Classification: Internal Page 7 of 74

8 1. Introduction 1.1. Scope of the document The scopes of this document are to report of the procedures being used for executing the project and describe the analysis process and achieved results Team composition The project was created based on the commitment from the consortium members to execute the actual work as well as the expert advisory group to support with reviewing and commenting to assure that the objectives are met. The Consortium members are AVTECH, LFV Group, Novair, Airbus and Egis Avia. The Expert Advisory Group (EAG) members are GE Aviation, Thales Avionics, Thales Air Systems and Scandinavian Airlines Systems (SAS) Revisions This document is updated by AVTECH, Sweden. Comments or questions shall be sent to: Christer Forsberg Classification: Internal Page 8 of 74

9 2. Project Objectives 2.1. High-level Objectives The MINT project had the overall objective to demonstrate minimum CO 2 in the TMA during descent by addressing efficiency aspects related to both lateral and vertical flight path. Eleven (11) demonstration flights into Runway 01R at Stockholm Arlanda airport have been carried out (10 to support performance analysis during revenue flights, and 1 dedicated to support dissemination activities, with invited media and industry representatives) Demonstration objectives The flights have been conducted with Airbus A321 aircraft with the most recent and advanced FMS available. The optimization of the vertical profile has been done by the aircraft FMS and supported by NowCast wind uplinks in order to enable the aircraft to reach state-of-art performance from an efficiency perspective. The lateral track has been addressed by utilizing the aircraft RNP capabilities in order to enable minimum track and noise abatement within the TMA. The longitudinal control of the flight should be addressed by flying the descent and approach in managed mode thereby letting the aircraft FMS control profile and speed. To assure a minimization of CO 2 the flights were flown with a low Cost Index 2 (below 8). Five of the eleven flights were constrained with a time requirement, or Control Time of Arrival (CTA), to a waypoint near FL100 during descent. This showed the feasibility to combine Continuous Descent Approaches (CDA) with time based operation as foreseen in the SESAR operational concept Quantify the potential objectives The efficiency in terms of fuel burn, and thereby CO 2, was analyzed using aircraft recorded data and reports the optimum fuel use in relation to reference samples from normal operations. These references are presented as an average fuel burn for flights in the TMA today (divided into groups of entry points and runway in use). All data were normalized in terms of different parameters such as wind, aircraft mass, ISA deviation, and cruise altitude in order to be comparable. For the flights that were assigned a CTA the accuracy of meeting the time i.e. the CTA performance of the aircraft was analyzed. 2 Cost Index (CI) is an airline individually chosen parameter that sets the relationship between cost of time and cost of fuel in the realization phase of the flight mission. When operating with CI=0, priority is given to fly with minimum fuel usage. Classification: Internal Page 9 of 74

10 2.4. Dissemination Objectives The main MINT dissemination objective was obtained by conducting one (1) seminar that presented the findings and demonstrated the aircraft RNP capability and optimized performance during a demonstration flight on November 25 th Classification: Internal Page 10 of 74

11 3. Trial Preparations and Execution 3.1. Flight trial preparations Getting the authority s acceptance The approach procedure used during the MINT trials is the result of several years of work and cooperation between LFV, Stockholm Arlanda Airport, SAS, and the Swedish Transport Agency (TS) (the Swedish regulator) among others. When the MINT project was initiated, this approach procedure was identified as a candidate for carrying out the 11 flight trials. However, when the project began, the Novair configuration of their Airbus aircraft was not certified to fly RNP-AR procedures and the procedure itself had not been approved by TS. A dialog was begun between Airbus, Novair, LFV and TS to obtain permission to carry out the 11 trials on a temporary basis. On the basis of a number of conditions, permission was granted to carry out the trials. These included: 1. Trials should only be carried out in good weather. In particular, a feet cloud base were required, 10 km visibility and no convective weather in the vicinity of the airport. 2. Trials should only be carried out when a test leader was present in the ATCC. There were three test leaders who were involved in the project and were well known by the flight crews and ATC. Their presence was very beneficial to ensure that the MINT flights could be carried out. 3. Trials should only be carried out with a flight crew who had been prepared and briefed prior to the flight. Next to this most MINT flight crews had also flown the approach in a simulator and knew what to expect and how to carry out the RNP procedure. 4. Trials should not be carried out while approaches were carried out to the parallel Runway 01L. This was a limit set by the TS. When that permission had been received, the procedure of how to carry out the trials was agreed upon between Novair, Stockholm ATCC and the Arlanda Control Tower. A number of safety cases were conducted based on these procedures, see Annex E. When all required safety criteria was fulfilled, management personnel responsible for operations at the tower and ATCC granted final approval for the trials. The developed instrument approach procedure that was accepted for the MINT flights can be seen in Annex A. In order to support the acceptance process for Novair to fly the new RNP-AR 0.3 approach in the frame of the MINT trials, airline and aircraft preparations were needed as described in chapter and chapter Airbus also organized a flight simulator session in Toulouse, where the flyability of this procedure was witnessed by TS, LFV, Novair and AVTECH. Various scenarios were flown with various wind conditions, with and without Required Time of Arrival (RTA) constraints. In addition, earlier simulation sessions by Airbus provided valuable input to the instrument approach procedure design itself, notably with respect to adequate speed limits, which had been taken into account in the final published version of the RNP procedure Aircraft preparations In order to enable Novair to fly the new RNP-AR 0.3 procedure into Stockholm Arlanda Runway 01R in a correct manner, Airbus provided the latest version of the Thales FMS (Release 1A) which was then installed by Novair through software upgrade. Classification: Internal Page 11 of 74

12 For Thales FMS, the R1A is the first version which is certified for RNP-AR approaches. The installation of this latest FMS version onboard Novair A321 was a pre-requisite for the MINT trials because the new noise abatement procedure at Stockholm is classified AR (Authorization Required) due to its left- then right-turn with constant radius after the Final Approach Fix. All Airbus aircraft leaving the assembly line have RNP-AR 0.3 potential. By far the majority of the Airbus fleet in service has the potential to be upgraded to RNP-AR 0.3. The extent of the upgrade required depends on the actual equipment level of the individual aircraft and can range from simple software loading to important avionics hardware upgrade. For an individual airline, the business case is mainly driven by the equipment level of its fleet and the benefit potential of the specific RNP-AR 0.3 procedure. For the purpose of data collection and in order to capture relevant data not recordable by Quick Access Recorder (QAR), Airbus provided the programming of the Digital Aircraft Condition Monitoring System Recorder (DAR), which then was loaded by Novair Airline preparation Novair coordinated the flight procedures with support from Airbus. When the trials commenced, Novair aircraft were not certified for RNP-AR 0.3 operation in accordance with the applicable EASA reference document AMC 20-26, which became effective on the 23 rd of December 2009 and no public flight crew publication were available. Novair prepared specific material for its flight crews regarding the MINT flights, based on best practices about how to conduct RNP-AR approaches with substantial support of Airbus. Novair trained some of its flight crews in conducting the RNP-AR procedure for Runway 01R in the flight simulator at Oxford Flight Academy, Stockholm prior to the test flights. In addition, the Navigation Data Base of the FMS was programmed in a short period of time for the flights, based on the approved procedure in May Development of the Operational and Data Collection procedure During the preparation phase the following procedure was developed. Before the flight Before the flight initial coordination should be made between Novair and the trial leader at ATCC in Stockholm. All candidate flights should have a flight crew that is informed and prepared for the operational procedure for the RNP-AR flight. Approximately 60 minutes before landing When the flight passes the Swedish FIR boundary and enters the sector handled by the ATCC in Malmö, coordination with Stockholm ATCC shall start and the flight crews should then, if still possible to run the trial, update the FMS with the new STAR. Novair AOC request FMS Performance Report to get aircraft mass, cost index and cruise altitude data as input to NowCast Wind uplink system. NowCast Wind uplink to the aircraft via ACARS (prior to ToD). The NowCast Wind uplink system (AVENTUS NowCast System) winds are defined for 5 different altitudes. Flight crews insert these winds into the aircrafts FMS manually. Classification: Internal Page 12 of 74

13 Flight crew manually update the FMS with applicable data such as ground temperature, ground wind, barometric pressure (QNH) and flap configuration for landing prior the CTA is defined. For CTA selected flights the flight crews insert the MINT-waypoint 3 and communicate the aircraft calculated ETA for this waypoint to Stockholm ATCC. RTA is assigned via voice communication by Stockholm ATCC and will be close to the ETA in order only to demonstrate RTA capability without penalizing the aircraft with e.g. delay in a longitudinal holding pattern. Once activated the RTA function of the FMS will adjust the speed accordingly until passing the MINT waypoint. At ToD At the FMS computed ToD, the crew initiates the descent. Descent altitude clearance should be given in time to avoid late descent engagement. During Descent Flight crews shall allow the aircraft to be flown by the FMS until the latter stage of the approach. Descent altitude clearances should be given in time to avoid cases where aircraft need to level off due to late descent clearance. Inbound clearance is confirmed. Within Stockholm TMA Control Flight crew judges that applicable weather minima are complied with. After Landing Flight crews print a FMS Post Flight Report to be used in the post flight analysis. Novair collects the QAR data for the analysis, DAR data and flight crews notes and sends the QAR data and flight crew notes to AVTECH and the DAR data to Airbus Special preparations for the Dissemination event flight The demonstration flight, which was scheduled to take place in November, for the dissemination event required specific coordination and planning. The coordination and planning for this flight started after the completion of the initial 10 flights. In general, weather in Sweden in November is not conducive to conducting trials that require good weather. There are generally days with low cloud base and/or precipitation and the number of hours of sunlight is limited due to the late rising/early setting of the sun. To mitigate possible effects of this, contact was taken with the TS to further reduce the weather restrictions that had been used during the initial 10 flights. As stated in Section 3.1.1, the initial conditions were set to feet cloud base and 10 km visibility. However, for the final flight, this restriction was lowered to feet cloud base and 5 km visibility. 3 The MINT-waypoint is a defined waypoint near FL100 that is stored in the aircraft to be used for the CTA trials. Classification: Internal Page 13 of 74

14 Extensive coordination with the Arlanda tower took place which included discussions about how to handle sub-optimal weather conditions, other traffic, eventual go-around handling etc. Three main scenarios were identified when planning the final flight: 1. Good weather and visibility and northerly or no winds. In this case, the trial could be carried out without larger deviation from normal operations. 2. Low cloud base and/or snow/rain In case the cloud base was well under the allowed limit or there was heavy snow, fog or rain, it was decided to continue the trial, but the flight would not fly the RNP-AR approach and would instead fly a P-RNAV (RNAV 1) STAR to the current runway in use. 3. Strong southerly winds When strong southerly winds were present which exceeded the aircraft s maximum allowable tailwind, it was decided to fly the RNP-AR approach to Runway 01R down to around feet. At that point, the aircraft would level out, fly over the airport and then be vectored for an approach to the current runway. The MINT flight was given the highest possible priority for normal aircraft in the Stockholm TMA. The only flights that would have higher priority would be emergency flights or flights with a head of state. Once these scenarios had been established, a series of meetings were carried out with representatives from Novair, ATC (tower and ATCC) and the test leaders from the MINT project where fallback and go-around procedures were agreed upon. In terms of where the aircraft would fly, it was decided that after take off, the plane should fly towards northern Sweden where the traffic levels are generally lower than in the south or east. The approach would then take place towards the HAMMAR VOR, which is also acting as a TMA entry point. This was the first time a trial had taken place towards this entry point and this approach is considered the most environmentally efficient because of a large number of saved track miles compared to a traditional ILS approach Flight Demonstration Scenario Set-up Internally, the trial leader began by checking the weather information by consulting the meteorologist at Arlanda airport to get the most up to date prognoses for the day. In case the weather was 100 % certain to not be amenable to trials, the flight crews were contacted immediately and the trial cancelled. Otherwise, the flight crews and test leader maintained contact during the flight to discuss the local conditions at Arlanda. The test leader also coordinated with the local watch supervisor in the Stockholm ATCC and got his/her feedback about how the traffic situation and weather would develop. In addition, if any extraordinary activities were planned for the day such as runway closures, or reduced rates, they were discussed. The test leader also talked to the tower watch supervisor at this point to discuss if there were any airport related activities that would affect the trials. Finally coordination was also made with the person responsible for the AVENTUS who coordinated with Novair AOC. Classification: Internal Page 14 of 74

15 Execution Once the set-up process had identified a possibility of executing the flight, the process started as described in the operational procedure specification, chapter Data collection As described already in the chapter describing the operational procedure and data collection description, section 3.1.4, a number of data in various formats was collected after each flight. These were mainly: QAR DAR Controller log Post-Flights reports MAESTRO log Interviews notes The use of the different collected logs is described in chapter Performing the trials The first flight was carried out on 16 th of June 2009 when Novair flight NVR352, with Captain Henrik Ekstrand, from Zakinthos in Greece, which received clearance to fly the RNP approach to Arlanda s Runway 01R via the Trosa VOR TMA entry point. The flight crews reported very positive aircraft performance and that the arrival and approach went as expected but they had to stay above crossing traffic at one point so the vertical profile was less than optimal and for this reason, the flight was excluded from the performance analysis. Between 17 th of July and 14 th of August 2009, 10 additional RNP flights were carried out and the actual data was logged to be used in the analysis. Four of the 10 flights flew with the CTA constraint as described in chapter and chapter 5.4. In parallel to the 10 successful flights, 12 flights were initiated and cancelled at a late stage (in every case due to convective weather activity, low ceiling or southerly winds) The demonstration flight was successfully carried out on the 25 th of November This flight also was flown with a CTA constraint but via the northern Hammar VOR TMA entry point. Classification: Internal Page 15 of 74

16 4. Analysis The analysis has been made on the collected information described above by AVTECH who has used already developed tools and systems. All the results has been circulated internally in the project and discussed when necessary Efficiency analysis The efficiency analysis is made based on comparison between data logs from flights conducted before the introduction of the MINT procedure and the MINT flights. In order to be able to compare the flight logs a normalization process is used that is part of the AVTECH Efficiency Analyzer tool that has been used. In order to fully understand the benefit of the normalization process, we take two flights as an example: Flight A and flight B have exactly the same states and the same trajectories (same mass, altitude and ISA condition). However, flight B has a headwind during the final approach which makes the flight more fuel demanding. If we only consider the real fuel burn, flight A is more efficient compared to flight B. The wind directly impacts efficiency and biases the analysis. The fuel burn normalization aims to get the same fuel efficiency for these two flights whatever the wind magnitude, differences in mass, ISA deviation and cruise altitude. The different parts of the normalization process are described further in Annex B but in short the normalization starts from a given reference distance from the runway threshold in order to have comparable analysis between each flight. Then the process starts by normalising the wind, temperature, mass and altitude. The wind normalization finds the impact of the effective wind to the true air speed and the ToD location. The temperature normalization estimates the difference in fuel consumption between the real flight and what it should have burned in ISA condition. The mass normalization calculates the vertical profile of a reference flight at a given mass depending on the characteristics of the flight and then estimates the difference in fuel consumption between the flights due to different ToD locations. The altitude normalization, finally, calculates the fuel the flight should have used if it were at a given altitude during cruise and estimates the difference in fuel consumption due to variation in ToD locations Wind uplink accuracy analysis Based on the selected winds that were chosen by the AVTECH AVENTUS NowCast system and the knowledge of how the FMS uses these winds to predict the entire vertical wind situation, a comparison can be made between the predicted and actual wind profile. In order to validate the possible improvement from the AVTECH AVENTUS NowCast System and the current system the operational system selected winds was also logged and analyzed the same way Navigation accuracy Navigation accuracy is analyzed for all the logged positions during the arrival and approach. It is separated in the different phases that have different requirements RTA accuracy For five flights a CTA was assigned. In the aircraft this in managed through the use of the aircraft RTA functionality. The analysis of the RTA accuracy is simply made by comparing the controller/flight crew notes of the assigned time with the time when the aircraft actually passes the waypoint used for the time constraint. Classification: Internal Page 16 of 74

17 Noise analysis The noise distribution is based on the INM tool version which is integrated in the AVTECH Environmental Analyzer tool. It uses the logged N 1 parameter (the rotor speed of the engine low pressure compressor axis) together with position and altitude information to calculate the ground noise distribution Collection of future ground system functionality requirements This analysis has been made by Egis Avia based on discussions with people involved in the trial and will simply be reported as concluding outcomes from these dialogues in this report, se section 5.6. The main question asked is What updates can be made in the supporting ground systems in order to support this optimized arrival procedure to be offered also during more dense traffic situations? The focus has been made on requirements related to the destination airport arrival sequencing tool, the Arrival Manager (AMAN). 4 The Integrated Noise Model (INM) is a computer model that evaluates aircraft noise impacts in the vicinity of airports. It is developed based on the algorithm and framework from SAE (Society of Automotive Engineers) AIR 1845 standard, which used Noise-Power-Distance (NPD) data to estimate noise accounting for specific operation mode, thrust setting, and source-receiver geometry, acoustic directivity and other environmental factors. Classification: Internal Page 17 of 74

18 5. Results 5.1. Efficiency The efficiency analysis of the MINT flights in relation to three reference logs into Runway 01R from the same direction show savings in the order of 165 kg of fuel, corresponding to 520 kg of CO 2, when comparing with 3 baseline flights 5 into the same runway. The results are collected from the following numbers (all flights are analyzed with the IRP 140 NM from the runway. See the definition of IRP in Annex B): Average MINT flights (11 flights): Actual Fuel Burn: 592 kg Normalized Fuel Burn: 509 kg Track Mile: NM Average ILS flights into Runway 01R (3 flights): Actual Fuel Burn: 706 kg Normalized Fuel Burn: 674 kg Track Miles: NM Two baseline flights to Runway 19 were also analyzed (this is interesting for comparison to time periods after 23:00 local time, when Runway 01R is unavailable for landing and Runway 19L/R can be used as the alternative) from the same IRP point as for the 01R threshold. This analysis shows the possible benefit if being able to open the 01R runway after 23:00 local time by operating the RNP procedure, avoiding the noise sensitive areas. ILS flight into 19L Actual Fuel Burn: kg Normalized Fuel Burn: 841 kg Track miles: NM ILS flight into 19R: Actual Fuel Burn: kg Normalized Fuel Burn: 715 kg Track miles: 174 NM This means that the RNP01R procedure saves in the order of kg of fuel for flights where Runway 01R can be used instead of Runway 19L/R. Based on the procedure of identifying an IRP for each flight, as described in Annex B, it is also possible to use flights from different directions flying into different runways as baseline flights. Below one can see the drawings of all the MINT as well as 28 logs received serving as baseline flights from Novair. In Table 1 the corresponding results can be seen. The IRP was chosen 140 NM away from the runway threshold. 5 Runway 01R is used in a limited quantity by Novair and the number of baseline logs is limited to only 3. Classification: Internal Page 18 of 74

19 Figure 1. All 10 MINT flights as well as all 28 baseline flights. Table 1. Analysis of MINT flights compared to other baseline flights into other runways. Based on this we can conclude that the track into Runway 01L for these Novair flights seems very efficient, even shorter than the RNP procedure. The RNP procedure into Runway 01R won t be the shortest since it circumnavigates noise sensitive areas. Even if the 01L approach is more straight-in (and thus has less track miles) we still burn some 30 kg less fuel with the MINT procedure due to a more efficient vertical profile. Classification: Internal Page 19 of 74

20 For the MINT project, the trials flew on average 3 NM shorter approaches compared to the average of those of the baseline flights that flew to the same runway. Their fuel consumption per NM in cruise was equal to 5.88 kg/nm. It means that the fuel saving due to shorter lateral track was equal to 20 kg of fuel. By deducing the fuel savings linked to track miles, we obtain the fuel saving due to a better vertical profile. In average, the MINT trials saved 145 kg of fuel per approach due to an optimized descent profile and where the flight crews had a high predictability in the lateral flight trajectory. For the demonstration flight a separate analysis was made since it was the only flight arriving from the north into Arlanda as previously stated. No baseline logs are available for this individual approach (from the north landing on Runway 01R). So in order to get an estimation of the savings compared to a traditional ILS flight coming from the north a simulation with the TYDe A321 simulator 6 was made at AVTECH. The procedure in this case was a CDA approach coming from the north down to a level segment at feet for the ILS intercept (about 15 NM from the runway) 7, i.e. a rather good performing traditional approach representing a low traffic density scenario. The results from comparing the demonstration flight RNP approach with the simulated ILS procedure it was given that the ILS approach is 14 NM longer and that fuel saving for the RNP flight was 333 kg (based on normalized values) Navigation Lateral analysis The Performance Based Navigation analysis looks at the lateral accuracy of the aircraft during the approach. The analysis is made from two different points of view: - RNP 0.3, i.e. from the waypoint SA693 (IAF) down to SA From the FAF (waypoint SA695) where the RNP-AR criteria starts. All MINT flights are included into the analysis. The samples used in the analysis are all with auto-pilot and navigation mode engaged. Every sample corresponds to a one second recording from the onboard equipment. The picture below (Figure 2) shows the tracks of the MINT flights flying the RNP-AR procedure along the different waypoints. As can be seen one track is slightly deviating after passing SA692. This was due to convective weather activity and they had to circumnavigate momentarily. Since the lateral analysis doesn t start until after SA693 this is not affecting the lateral analysis. 6 TYDe is a Thales developed Test at Your Desk simulator that is built around the actual FMS hardware and software. 7 At Arlanda airport, a feet glideslope intercept of the ILS is used for arrivals into Runway 01R due to lowering the noise in the suburb Upplands-Väsby south of the airport. For other runways the glideslope of the ILS is intercepted at feet. Classification: Internal Page 20 of 74

21 Figure 2: MINT Flights RNP Approaches When looking at the lateral deviation from SA 693 to SA 695 (Figure 3) the best fit normal distribution has a mean value of µ = NM (-9.9 m) and a standard deviation of σ = NM (40.8 m). The left and right maximum values are respectively NM ( m) and NM (235.2 m). Classification: Internal Page 21 of 74

22 Figure 3: Lateral Deviation below feet When analyzing the lateral deviation after SA695 (Figure 4) the best fit normal distribution has a mean value of µ = NM (-13.1 m) and a standard deviation of σ = NM (21.2 m). The left and right maximum values are respectively NM (-83.2 m) and NM (28.9 m). EUROCAE/RTCA ED-75/DO-236 MASPS (Minimum aviation system performance standards: Required Navigation Performance for Area Navigation) document gives the lateral requirement for RNP accuracy: The aircraft position error must be less than the RNP value for 95 % of the flying time. Given the normal distribution of the cross track error, the 95 % of the time corresponds to the double of the standard deviation. The accuracy of the MINT flight is thus equal to NM from the FAF when the RNP is 0.3 NM so as can be seen the MINT flight accuracy completes the requirement. Classification: Internal Page 22 of 74

23 Figure 4: Lateral Deviation from SASA695 (FAF) Vertical profile The vertical profiles of the MINT flights have been compared with a sample of the reference flights as can be seen in Figure 5. The reference flights are fairly good which is expected since Novair flights normally arrive during low traffic density periods and as such they receive efficient descent clearances. This indicates that potential savings identified in the project could be seen as a possibility also for smaller airports with already efficient arrival procedures. As shown earlier the MINT approach saved approximately 30 kg of fuel compared to the standard approach to Arlanda Runway 01L even though it was 5 NM longer (due to the circumnavigation of the noise sensitive area). Airbus FMS is designed to the objective to fly "hands-off" fully managed idle descent by decelerating during short horizontal segments and descending with constant speeds. In the case of the MINT trials, the flight crew further optimized the vertical profile through partly avoiding the horizontal deceleration segments by tuning the vertical speed; no significant impact on the workload was observed. Ongoing Airbus autopilot developments will make this manual fine tuning task unnecessary and will even provide advice to the flight crew at which moment to set the configurations (slats/flaps) for an optimized continuous descent approach. Classification: Internal Page 23 of 74

24 Figure 5. Comparison between the 10 MINT flights and a sample of the reference flights Wind uplink The Airbus FMS can receive five wind levels with forecast wind information. The wind uplinks for the MINT flights were provided by the AVTECH AVENTUS NowCast System. During traditional flights, and as such during the reference flights, the airline has uplinked winds provided by the flight planning system, in this case the RODOS system. The difference between the systems are that the AVENTUS selected winds are collected along the aircraft s trajectory and the five levels are selected dynamically to minimize wind errors and maximize the onboard equipment ability to predict the total wind profile instead of selecting winds only based on fixed levels and from positions that is not necessarily along the trajectory. During the flight trials both the values provided by the AVENTUS system as well as the flight planning system were logged. The FMS wind component prediction, based on the selected values from AVENTUS and the flight planning system respectively is then compared to the actual wind components and the correspondence is calculated. From the MINT flights it has been identified that the average AVENTUS correspondence is 83% compared to the actual winds and if the traditional flight planning system would have been used the corresponding average would have been 65%. The results from each individual flight could also be seen in the table below. Flight NVR392 17JUL NVR292 3AUG NVR352 4AUG NVR422 5AUG NVR242 6AUG NVR392 7AUG NVR362 9AUG NVR392 14AUG NVR NOV AVENTUS accuracy (%) 95,29 84,37 75,74 65,85 86,59 78,94 82,24 95,73 78,24 Classification: Internal Page 24 of 74

25 Flight planning system accuracy (%) 86,22 70,38 68,12 49,56 60,83 26,10 61,29 89,43 75,23 Table 2. Result of the FMS wind prediction correspondence to actual wind profiles based on inputs provided by AVENTUS and traditional flight planning system respectively. To illustrate the effect from the AVENTUS advantage of being able to select the levels for the five wind inputs dynamically the following figure can be analyzed. In the reference flights (in normal daily operation) the aircraft receives uplinks with descent wind information for five fixed selected flight levels; FL300, FL240, FL180, FL100 and FL050. In the example below the FMS misses the actual tailwind at around FL170 as well as some of the headwind below FL50. In the daily operation the Novair pilots then replaces the FL50 wind with the ground wind when available which should refine the wind profile slightly but at a late stage (usually after the ToD). Figure 6. FMS prediction results based on inputs from AVENTUS and flight planning system compared to actual winds RTA accuracy Four of the ten ordinary flights as well as the MINT demonstration flight were time constrained from the ground by a CTA to a specific waypoint (called MINT point ) located FL100. The ordinary trial flights used the MINT point whereas the demonstration flight used the TMA entry point HAMMAR VOR. The time constraints were given from the ground controller before the ToD, approximately 30 minutes before passing the point. Important to remember is that the CTA given to the crew was close to the aircraft estimated time of arrival for the specific point. Hence this procedure doesn t prove the aircraft capability to take a delay during en-route but merely indicate the aircraft capability to control to a time in combination to flying the efficient approach set-up. The following picture shows the location of the MINT point for the ordinary trials. Classification: Internal Page 25 of 74

26 Since the assigned times were close to the aircraft preferred times no difference has been identified in terms of resulting fuel burn. Hence, even with the RTA activated the aircraft has been able to fly with the speeds corresponding to an efficient flight. Large changes to the times would have required the RTA function in the FMS to change the descent speed (to a less optimal in regards to fuel) in order to prioritize meeting the time. Any analysis of the effect in fuel burn for larger changes in time for the aircraft would require a larger number of trials. Figure 7: MINT Waypoint location The time accuracy has been calculated from the logged information completed by the ground controller and from the on-board collected data. One gives the time constraints and the other gives the over-flown times on the MINT waypoint. The difference between the two times gives the RTA accuracy. The objective with the RTA procedure was mainly to demonstrate the aircraft capability to control to a time in combination to flying the efficient approach. Still we take the opportunity to analyze the RTA accuracy from these flights. Due to the very limited amount of flights it is important to remember that these RTA results are only indications of the aircraft RTA capability. With this in mind the results of the MINT flights was that the average time for the flights was 2.8 seconds on the late side, the standard deviation was 11.7 seconds and the average error for the flights was 8.4 seconds (i.e. the average based on the absolute time error values). Classification: Internal Page 26 of 74

27 As background information it should be stated that the present FMS specifications have a 30 second tolerance setting for the RTA functionality which means that the FMS will not make any changes to the speed before the estimated time for the RTA point is off compared to the assigned time with 30 seconds or more. As long as the estimated time is within 30 seconds from the assigned time it will see itself as on time Noise distribution effects The baseline for the noise issues into Arlanda is that the traditional approach is passing straight over the northern Stockholm suburb Upplands Väsby and almost also straight over the hospital Lövenströmska Sjukhuset. The noise level at the hospital is the main reason for Arlanda not using Runway 01R after 23:00. The results from the RNP-AR procedure into Runway 01R show according to the noise plots that Upplands Väsby suburb should receive a large reduction of the noise in the area. For the hospital the noise reduction could be up to 10dB(A) due to the circumnavigating maneuver. See Figure 8. Figure 8. Noise impact for the suburb Upplands Väsby as well as the hospital Lövenströmska Comparing one of the flights during the trials and one simulated traditional approach the results looks like in the Figure 9 below. In Figure 10, the inhabited areas can be seen. Classification: Internal Page 27 of 74

28 Figure 9: Noise footprint comparing ILS and RNP-AR procedures into Runway 01R. Classification: Internal Page 28 of 74

29 Figure 10. Inhabited areas compared to ILS and RNP-AR procedures into Runway 01R The noise distribution for the demonstration flight, the results from the real flight as well as results from a simulated traditional track is included below. Classification: Internal Page 29 of 74

30 Figure 11. Noise distribution compared for demonstration flight and a simulated traditional flight path Collection of Future ground system functionality requirements The Arrival Manager system, so-called AMAN, is one piece of the arrival system that encompasses pilots, air traffic controllers, airspace organisation, air systems, air-ground communication means and ground systems. At present, there is no SESAR service defined for arrival management. As a result in the project MINT, the partners propose some guidelines for setting a first vision of the provision of an arrival service. This arrival service is specified in various possible services levels detailed in the table bellow depending on the air and ground equipments. The correspondence between the listed arrival service levels and the SESAR roadmap needs to be discussed in other projects and the actual table is a proposal for feeding other projects. The list below is the result of interviews and comments from people that has been involved in the execution phase of the Mint flight trials. Classification: Internal Page 30 of 74

31 Service Name Description Comments Technical enablers Level L0 Basic situation Arrival service is only provided according to working methods. No specific tool is used to manage arrival flow: the arrival information is not computed and disseminated to the controllers, the flight This is the current situation in many airports that are not yet equipped by an Arrival Management System. The arrivals are controlled safely but during peak hours the use of holding patterns are VHF, radar crews is informed of the arrival instruction needed impacting the overall capacity, step by step by each control sector. efficiency, predictability and the environment. Vectoring the aircraft makes pressure on pilots that must cope with unpredictable trajectories. L1 Automation on the ground side Basic AMAN L2 Advanced automation ground side on the L3 Using aircraft trajectory The arrival service is supported by an AMAN tool and by working methods. The AMAN tool computes and updates an arrival sequence which is disseminated to all concerned controllers who works in coordination to achieve the arrival plan. The arrival sequence can be provided to the airport and aircraft operators via a dedicated connexion. The arrival management function is completed to be capable of handling mixed trajectories such as RNP, conventional procedures, vectoring working methods Controllers can request flight crews via voice (VHF) to manage a time constraint on a point. Taking advantage of the capabilities of new generation of FMS and of communication means, the aircraft trajectory is taken into account to build and update the arrival sequence. It is a first step to open the service to aircraft or airlines. Some busy airports are equipped in the world and deliver such services. Such organisation prevents the use of holding pattern: the benefits are significant in term of capacity, efficiency (fuel and gas emissions savings), predictability, improved coordination and less workload RNP procedures are emerging at some airports in the world but the existing AMAN tools need to be adapted to cope with mixed traffic in order to make controller s workload acceptable, to maintain high level of capacity and safety. Downloading the aircraft trajectory can be envisaged by several means such as ADS-B, ATN, AOC. The service is independent of the technology used but requires a data link infrastructure. The service should benefit to the Classification: Internal Page 31 of 74 Advanced AMAN and ground TP. Aircraft RNP and RTA capable. Basic data link means. FMS exporting the trajectory. Ground system handling trajectory.

32 Service Level Name Description Comments Technical enablers L4 Providing aircraft with arrival information L5 Contracting the arrival segment of the business trajectory The datalink infrastructure is used to uplink arrival information to the collaborative aircraft or airlines. The arrival information includes the landing runway, the STAR and approach trajectory, a time information over a point of the route. The pilots use FMS to make decision to accept or request for renegotiation. The full cooperation between flight crews and controllers is supported by advanced tool that prepare an efficient contract committed by the operators. The arrival contract includes an arrival trajectory defined laterally, vertically and a time constraint on a metering point possibly completed by a speed constraint in case of necessity. collaborative aircraft that are sequenced with accurate information as soon as they provide their trajectory. AOC network of CDM organisation could be used for this service because it is limited to an information level. AMAN and communication functions must be adapted to manage the dissemination and the update of the information according to Service Level Agreement. This ultimate level of service requires the commitment of all the sectors/controllers concerned by the arrival trajectory (At least En route approach and Tower controllers) and of the flight crew. Table 3. Various possible services levels depending on air and ground equipments Basic datalink means Advanced datalink infrastructure advanced ground tools to manage contracts (air ground and ground ground) => SWIM Classification: Internal Page 32 of 74

33 The intensive discussions during the project raised the need to harmonise the vocabulary and more specifically the definition of the various points of the trajectory during the descent. Three key points were identified that have different functions but can be merged in a single named/published point: feeder fix: this point sets the transition between En Route and approach centres and is currently used by arrival manager to enable a silent and efficient coordination between controllers. AMAN is using such point for displaying the sequence at this reference and algorithm are using the estimated time over the feeder fix to compute the arrival constraint at the runway threshold. This coordination point is also useful for flight crews to get the Estimated Approach Time that must be known to perform an approach in case of a radio failure. metering points : this point is the reference for applying the RTA/CTA, possibly with speed constraint when necessary. The location of the metering point depends on each arrival flow shall be adapted to the runway configuration and possibly to the kind of traffic (propeller/jet/jumbo). merging points : this navigation points are within the TMA and are used to merge various flows in a single one. For this point, fixed rates could be applied to prevent safe separations of the flows Results and observations from ATC 11 trials were carried out during the MINT project 10 ordinary and the demo flight. In addition to these flights, 12 flights were prepared and cancelled at a late stage (in every case due to thunderstorm activities, low ceiling or southerly winds). During the initial trials, the controllers were given a map of the expected lateral profile of the aircraft, but as seen in Figure 12, the lateral profile of the aircraft was not present on the radar screen. This made judging merging situations difficult. After the trials were continued, a plot was made available on the radar screen which makes judgement of the lateral profile much easier. Classification: Internal Page 33 of 74

34 Figure 12. Radar screen showing the MINT flight in green on approach to Runway 01R Another factor which made judging merging situations even more difficult was that MINT flights flew relatively slowly and had different vertical profiles compared to other aircraft flying CDAs. Figure 13 shows the first MINT flight. This flight had one other aircraft which needed to be merged with. When it entered the Stockholm TMA, it seemed as if the MINT flight would arrive well before the other (NDC506). However, as the situation developed, it became obvious that the NDC506 flight would arrive first. In the end, NDC506 landed six minutes prior to the MINT flight. This pattern was repeated in almost all subsequent MINT trials. Classification: Internal Page 34 of 74

35 Figure 13. Radar plot showing first MINT flight and another flight (NDC506) There were difficulties carrying out the trials which can be grouped into difficulties due to preconditions required to carry out trials (i.e. weather limitations) and difficulties in judging the vertical and lateral profiles of the MINT flights and thus merging them with other traffic. ATC is used to handling exceptional situations such as emergency flights, late runway changes, etc. and therefore there was no operational difficulty for ATC to carry out these trials. During the summer, southerly winds prevailed, which meant that Runway 01R (normally used during northerly winds) was seldom the runway in use. In addition, Runway 01L/19R was closed during the entire summer due to maintenance which meant that it could not be used to relieve other traffic, meaning more complicated situations. To handle these situations, some MINT flights landed with tail winds while other aircraft waited to start on the opposite runways. As a consequence of the MINT trials, other aircraft are punished. In total, five aircraft were delayed a total of 19 minutes on the ground and six aircraft were delayed a total of 12 minutes in the air for the 11 completed MINT trials. Classification: Internal Page 35 of 74

36 6. Possible Deployment Scenario and Related Results RNP technology is available and is becoming more and more accepted. Once a procedure has been published and airlines have started using it, airlines can start saving fuel and reducing the load on the environment from the first flight, provided it is designed from the perspective of minimizing track miles. Airlines will also benefit from increased predictability and knowledge of how they will fly to the runway. One remaining problem is to provide a solution for mixed equipage scenarios during high traffic density periods. With today s available ground support tools, air traffic controllers receive traffic that is not sufficiently sequenced when approaching the terminal manoeuvring area. In part, because of this, controllers need more possibilities to control traffic than what can be achieved with speed control along short, fixed routes. One main possibility for solving the sequencing issue is to continue the development of an AMAN system as elaborated on in Chapter 5.6. For shorter flights and when airports operate in mixed mode, it is expected that also departures will be required to be coordinated via use of departure management systems which includes cooperation between arrival and departure managers. For Arlanda the developed procedure into Runway 01R will go into operation during low traffic periods starting in January 2010 and further RNP procedures will be created in the near future for other runways and other Swedish airports. For Arlanda the results from the RNP procedure in terms of reduced noise for sensitive areas will also be analyzed further with the ambition that the limiting restrictions can be removed. For Arlanda the noise levels at the hospital Lövenströmska limits the airport to operate the 01R runway after 23:00. The initial result within the project is that the noise the hospital is exposed to could get be greatly reduced due to the RNP-AR procedure. Many flights each year need to circumnavigate the airport due to the noise regulations and if the new procedure could allow some of these aircraft to land on Runway 01R when preferred instead of circumnavigate the airport to Runway 19L/R the project analysis has identified the potential saving in the order of 300 kg of fuel per flight corresponding to almost 1 metric ton of CO 2. For Sweden, RNP development is part of a Performance Based Navigation (PBN) study which has been performed by LFV in accordance with the ICAO 36 th General Assembly Resolution A36-23, requiring its member states to submit a national PBN implementation plan. Part of this study includes a roadmap for developing and using RNP into many of the Swedish airports. Development of methods and tools to be able to use short RNP procedures in more dense traffic will also take place. At this stage, it is difficult to say when this can be achieved. For smaller airports the savings will come mainly from the reduction of track miles when approaching the runway from the opposite direction. In order to get an indication of potential savings from the RNP-AR technology seen in a European perspective the following elaboration is made. For the shorter term implementation timeframe, small and medium sized airports should be the one with the easiest implementation possibilities. The MINT trials indicate that RNP operations could be used today during low traffic and more coordination and support tools will be required to introduce such procedures into higher traffic density. For choosing a possible deployment scenario it is assumed that the short term target airports will exclude the largest airports due to their longer peak periods. From the Eurocontrol STATFOR Trends in Air Traffic vol 3, 2007 the number of airports in Europe are identified. Classification: Internal Page 36 of 74

37 Figure 14. Airports by size and number of known runways. Summarizing all the airports with less than departures (and thereby arrivals) per year gives us a total of 507. If deducting the smallest airports (with less than departures per year) the sum ends up at 224. The approximate number of departures for these airports is if assuming the average amount of departures for each interval. In the same source the average percentage of jet aircraft for each airport size category is presented and assuming that 25% of the jet aircraft could be RNP-AR approved and arrive in a low traffic density periods results in flights per year that should possibly have the potential to fly an approach as during the MINT flights. If assuming further that efficient RNP procedures could be implemented for half of these airports it gives flights per year. Finally, with the assumption that the fuel saving for each flight will be 140 kg, as for the average compared to all baseline flights into Arlanda, see Table 1, this equals to savings of over metric tons of fuel and over metric tons of CO 2 per year. With IATA fuel price of December 2009 (672 $ per metric ton) this would equal to almost 50 Million $ per year. Classification: Internal Page 37 of 74

38 7. Summary of Dissemination Activities Avionics 09 During March in Amsterdam, AVTECH reported as speaker in the Future Air Traffic Management conference session. The presentation addressed Performance Based Operations, Building Blocks of SESAR and NextGen Implemented and Used by ANSP s and Airlines Reducing Emissions. MINT was part of the presentation ATC Global 09 Partners with a stand at the exhibition provided a MINT project leaflet. Support SJU initiated AIRE dissemination event during the Paris Air Show 17 th of June 2009 at Le Bourget The MINT project leader presented the project and expected results based on simulations during the SESAR JU initiated AIRE dissemination event at the Le Bourget Paris Air Show in June Conduct a seminar event including a Demonstration flight This was the main dissemination event within the project. The event was carried out on the 25 th of November Around 100 guests participated on the dedicated demonstration flight that both departed and landed at Arlanda after a tour up north in Sweden. Among the visitors was for example the Swedish Transport State secretary Mr. Leif Zetterberg, Senior Vice President and Head of ATM from Airbus Mr. Eric Stefanello, representatives from SESAR JU and top management from numerous ANSPs (among others LFV Director General Lars Rekke), airlines (among others Novair Managing Director Anders Fred) and union organizations. The event was also covered by a number of media organizations, for example the Swedish television SVT. After the flight that was successfully executed by Novair Captain Henrik Ekstrand a seminar was arranged with two parts, one high-level part with presentations from managers representing different stakeholders, and one detailed session with presentations about the project and the collected results. See the invitation and agenda in Annex C. The media coverage from the related press release was very positive and extensive and the press release was quickly published on numerous web-news sites, for example (not including the consortium member s web-sites): GreenAirOnline.com EADS.com Yahoo Finance foxbusiness.com Business Wire Flightglobal.com Agence France Press - AFP ASD News Sesarju.eu All texts are attached in Annex F. Next to the Press Release the event also resulted in a video with interviews. Dialog with media, Flight International, Aviation Week etc. The MINT project has been covered in the following media during the year. Flight International Aviation Week ICAO Waypoints Newsletter GreenAir online Avionics Magazine Classification: Internal Page 38 of 74

39 Bulletins-electroniques.com Ny Teknik Dagens Nyheter All articles are attached in Annex F. Press Releases Three press releases have been published. See Annex D. Classification: Internal Page 39 of 74

40 8. Conclusions, Recommendations and Final Words The project has executed the 10 flights as well as the flight during the dissemination event successfully as intended. The data has been collected and analyzed based on a normalization methodology such that it allows different flights to be compared from an aircraft performance perspective. The results in this report are based on the limited number of flights and thereby they should be seen as indicative results for what can be achieved. The identified results from the flights is that, compared to the traditionally operated flights there are potential savings that can be made, seen from an aircraft perspective and not taking ATM related issues into account. The project identified 165 kg of potential savings of fuel for the 01R runway when arriving from the south and 140 kg potential saving if also including other directions and other runways to the baseline performance. It should be pointed out that the Novair flights most often arrives to Arlanda during low traffic density periods where they are offered fairly straight in approaches so this potential saving should indicated the minimum savings possible when utilizing optimized RNP and CDA techniques. Lateral navigation of the aircraft is excellent! For the ten flights during the RNP-AR segment the lateral standard deviation is excellent, measured to be 21.2 meters. In the case of the MINT trials, the flight crew optimized the vertical profile through partly avoiding the horizontal deceleration segments by tuning the vertical speed; no significant impact on the workload was observed. The result from this indicates a saving potential since the MINT flights had a lower consumption even though the baseline flights, for some comparison scenarios, even had a shorter track distance to fly from the reference point. The RNP procedure indicates also to be a very strong tool for addressing noise distribution problems by enabling circumnavigation of the areas. For Upplands Väsby, the suburb north of Stockholm, mentioned for having noise problems due to arriving traffic, the noise could be reduced largely with the defined RNP-AR procedure. Regarding Arlanda airport, the noise levels at the hospital Lövenströmska limits the airport to use Runway 01R after 23:00 local time. The initial result within the project is that the noise the hospital is exposed to could get be greatly reduced due to the RNP-AR procedure. Many flights each year need to circumnavigate the airport due to the noise regulations and if the new procedure could allow some of these aircraft to land on Runway 01R when preferred instead of circumnavigate the airport to Runway 19L/R the project analysis has identified the potential saving in the order of 300 kg of fuel per flight corresponding to almost 1 metric ton of CO 2. For Sweden, RNP development is part of a PBN study which has been performed by LFV. Part of this study includes a roadmap for developing and using RNP into many of the Swedish airports. Development of methods and tools to be able to use short RNP procedures in more dense traffic will also take place. At this stage, it is difficult to say when this can be achieved. Starting in January, 2010 the RNP procedure to Runway 01R will go into operation during low traffic at Arlanda. If looking on a rough implementation scenario seen in a European perspective and assuming that this technique can be implemented at small and medium sized airports and used as long as the traffic density is modest, the potential savings is still very interesting. Assuming half of the airports with to arrivals per year and assuming that 25% of the jet aircraft into these airports would be Classification: Internal Page 40 of 74

41 allowed to fly a short RNP-AR and CDA approach with the saving of 140 kg per flight this would add up to metric tons of CO 2 per year. The recommendations from the project are mainly: To continue to exploiting RNP and optimized descent techniques, initially into small and medium sized airports where the use can be initiated directly without added ground system support tools as foreseen for larger airports. This should be done at a new location with a different ATC environment and should preferably include the analysis of the flight from a runway to runway perspective. To support the identification and specification of requirements for an AMAN system, and other ground support tools, required for enabling the RNP implementation at larger airports. To continue the work with giving operational incentives to airlines to take advantage from potential quick wins resulting from RNP-AR 0.3 navigation capability (e.g. extended operating hours thanks to noise abatement). As with all flight trials, a large amount of coordination, preparation and communication is necessary between all actors. However, due to the fact that for the most part, the trials took place during extremely calm periods of traffic, it was not exceedingly difficult to allow the MINT flights to be carried out. As mentioned above in Section 5.7, there were possible difficulties identified for carrying out the trials which could be grouped into difficulties due to preconditions required to carry out trials (i.e. weather limitations) and difficulties in judging the vertical and lateral profiles of the MINT flights and thus merging them with other traffic. During these trials, however, there was no operational difficulty for ATC to carry out these trials. Finally it can be concluded that one obvious reason for succeeding with a project like this is a very engaged and deeply committed team. Classification: Internal Page 41 of 74

42 Annex A: Procedure for Arlanda RWY 01R Figure 15. RNAV (RNP-AR) procedure for Arlanda airport Runway 01R. Classification: Internal Page 42 of 74

43 Annex B. AVTECH Normalization Process Description provided for MINT This annex describes the process for normalizing the data and the background information about it. A. Initial Reference Point IRP identification The first point is identified thanks to a reference distance from the runway. As flights arrive from different directions with different procedures on different runways, it is important to build a process which can analyze data from a point identified as equal distance as the other flights for them to be comparable. For the MINT project, the reference distance is equal to 140 NM. The idea is to start the analysis from a point which is 140 NM far from the runway threshold by using the most direct route. The direct route is illustrated in Figure 16. RT F RT R C M T Trajectory Figure 16: Direct route identification The point T is part of the trajectory. The point M is tangent to the circle. The point F is 5 NM distant from the point RT (Runway Threshold). The turn radius of the circle is defined by considering a 30 bank angle and the lift equation during a stable turn. The objective is to find the first point T on the trajectory which is 140 NM distant from the RT by following the calculated direct route. This point will be called IRP for Initial Reference Point. By using this methodology of selecting a reference point the baseline results, acting as reference of the efficiency of today s operations, can actually be collected from flights flying into completely different runways as long as the operating scenario (type of STAR, traffic density etc) is comparable. Classification: Internal Page 43 of 74

44 B. ToD Identification If the flight had two different cruise segments (as seen in Figure 17), the ToD can be considered as the last point of the first cruise segment. However, in this case, the more realistic ToD is the last point of the second cruise segment since the first descent probably was due to crossing traffic or similar Altitude (ft) Time Figure 17 Vertical profile example This section describes how to select what can be considered as the cruise phase. In order to define the ToD, we need the following parameters which will be used for the ToD identification as detailed after. Cruise Altitude Limit (CAL): The CAL is the top altitude from which a stable segment is considered as a cruise segment. This parameter is used to identify if a stable segment is part of the descent or it is just a FL change due to performance or ATC purpose during the cruise phase. For the MINT study, the CAL is equal to feet. Minimum Stable Segment Distance (MSSD) Some of descent profiles make short stable segments above the CAL. Of course, we can not consider them as cruise segments. That s why we have to introduce a MSSD from which a stable segment is considered as a cruise segment (above the CAL). For MINT, the default value of the MDDS is 5 NM (i.e. a bit more than 1 minute of flying at 400 knots True Air Speed (TAS). Maximum Altitude Variation (MAV) Figure 18 illustrates how the process detects the end of a level segment. Classification: Internal Page 44 of 74

45 The process needs to know the maximum altitude variation from which it considers the flight in a descent or climb segment. If the altitude of a flight becomes below than the cruise altitude minus the MAV, the ToD is tagged as the last point at the cruise altitude. The default value for MINT is equal to 40 feet. Altitude MAV Figure 18 Descent Identification To identify the ToD point, we consider different kinds of trajectories as shown in Figure 19. For these following examples, we consider that every descent segment is considered bigger than the MAV and thus, the tool identifies a descent or climb phase. Flight Level MSSD CAL IRP 0 Distance to the runway threshold Flight F1 Flight F2 Flight F3 Flight F4 Flight F5 Flight F6 Classification: Internal Page 45 of 74

46 Figure 19 Descent profile examples 1. Flight F1 is the classic descent profile. The cruise phase is stable and the descent is continuous. 2. Flight F2 has a cruise segment and then, a lower one which is longer than the MSSD. The ToD is the last point of the lower cruise segment. 3. Flight F3 is similar to the flight F2. However the lower stable segment is under the CAL. This limit is defined in order to distinguish the stable segment between a cruise segment and an approach step. 4. Flight F4 has no data above the CAL. However there is no stable segment during the descent phase. That s why the ToD is the last point of the stable segment. 5. Flight F5 is already on a descent segment and has a stable segment below the CAL. The IRP and CAL parameters should have been set to avoid this problem. However the software has to consider this eventuality. For this kind of flight, the ToD is the first point of the study (IRP). If the stable segment were above the CAL, the ToD identification would be identical than flight F2. 6. Flight F6 has two stable segments during its descent. The first one is shorter than the MSSD. The second one is below the CAL. Thus the ToD of this flight is the last point of the cruise segment. Classification: Internal Page 46 of 74

47 C. Wind Normalization The objective is to eliminate the impact of the wind on the distance flown. From a fix geographical point, we have to find the distance delta the aircraft flies due to wind factor. GS dt TAS dt We dt Figure 20 Wind impact on the flown distance during dt seconds The Figure 20 introduces the relation between the ground speed (GS), the True Air Speed (TAS) and the effective wind (We). It shows the decomposition of flown distances due to the wind component. Therefore, for each dt seconds 8, the aircraft flies We dt NM more compared to a flight with no wind. N Heading GS N α W TAS α We Figure 21 Wind triangle Figure 21 represents the We in function of the GS, the TAS and the wind magnitude (W). The delta distance during dt between a no wind flight and a wind flight is: d( i) = We dt Then, by knowing the instant fuel flow and ground speed, we can, for each time step, evaluate the normalized fuel burn. When doing the normalization there will be a gap around the ToD. A headwind flight has a later ToD and a tailwind flight has an earlier one compared to a no-wind flight. The difference in ToD location is due to the air miles flown during the descent. For the normalization, we consider that the air distance during the descent stays equal for every flight. Thus we evaluate the difference between the air distance and the ground distance from the ToD until the runway threshold in order to know the fuel to subtract or to add depending on the value of the distance difference. 8 As the speeds are given in knots, the time given by these equations is in hours. However it is simpler to think in seconds instead of hours. Classification: Internal Page 47 of 74

48 Figure 22 represents the result taking into account the difference of the ToD. Figure 22: Actual and normalized fuel burn results during different wind conditions Classification: Internal Page 48 of 74

49 D. ISA Normalization The temperature has an effect on the fuel consumption during the approach because the flights do not fly at the same altitude. A flight flies indeed at the same pressure altitude but, as the temperature is not the same, the air density also and thus the isobaric levels are on a higher or lower altitude depending on the International Standard Atmosphere (ISA) deviation. Pressure ISA +0 Pressure ISA -10 Altitude Delta Pressure Altitude Altitude Height Pressure Altitude Altitude Height Mean Sea Level 1013 hpa Mean Sea Level 1013 hpa Figure 23 Temperature effect on altitude Figure 23 illustrates how the temperature impacts the altitude. With a lower temperature, an object (as the airport on the picture) seems to be higher in the 1013 hpa standard atmosphere. With a same cruise altitude setting, a flight flies lower with a negative ISA delta. Figure 24 illustrates the theoretical trajectories of different flights under different ISA conditions from the altitude and the pressure altitude perspectives. Classification: Internal Page 49 of 74

50 Altitude ISA +X Pressure Altitude ISA -X ISA -X Distance ISA +X Distance Figure 24 Trajectories with different ISA deviations A deviation in the temperature represents a deviation in the altitude by keeping the same pressure altitude. A flight flying under positive ISA condition starts its descent earlier but has a longer descent. We have here two cases: - ISA delta positive: o Higher altitude, earlier ToD and longer descent - ISA delta negative: o Lower altitude, later ToD and shorter descent However the ISA deviation is not constant during the flight as shown in Figure 25. That s why we cannot estimate the difference in fuel consumption due to the difference in location of the ToD points. Classification: Internal Page 50 of 74

51 ISA Temperature Deviaiton Novair First RNP Flight Pressure Altitude (ft) Thousands Flown distance (NM) Pressure Altitude ISA deviation Temperature ( ) Figure 25: Temperature ISA deviation during Novair A321 approach So in order to normalize the fuel burn for different ISA deviations, we use performance table (or BADA 9 database) in cruise and in descent. The idea is to change the current state defined by different parameters towards the normalized state. The Figure 26 shows the ISA normalization (made with BADA process only) for the Novair RNP flight flown on June, BADA (Base of Aircraft DAta) is an Aircraft Performance Model (APM) with corresponding database. BADA is being maintained and developed by the EUROCONTROL Experimental Centre (EEC). The Operations Performance Model of BADA defines, besides total energy model, the aircraft type, mass, flight envelope, aerodynamics, engine thrust and fuel consumption. The Airline Procedure Model defines the speeds that are to be used during the climb, cruise and descent flight phases. Classification: Internal Page 51 of 74

52 Figure 26: Novair RNP flight - ISA normalization made with BADA process The real fuel burn of the flight was 761 kg and is calculated to be 744 kg after normalization for temperature differences. Classification: Internal Page 52 of 74

53 E. Mass Normalization The mass has an impact on the fuel consumption during cruise because a heavier aircraft has to burn more fuel in order to keep the same TAS as a lighter aircraft and it also impacts the ToD location due to a lower rate of descent for a heavier aircraft than a lighter one as shown in Figure Pressure altitude Weight Simulations, 16/11/2008, AVTECH Analysis Pressure Altitude Flown Distance (NM) lbs lbs lbs lbs lbs lbs Figure 27: Different ToD for different mass Thus a heavier aircraft consumes more fuel during cruise but has a shorter cruise distance to fly compared to a lighter aircraft. The objective of the mass normalization is first to find the difference in fuel consumption during cruise and secondly to find the change in location of the ToD due to the difference in mass. We assume here that a flight and its normalized one keep the same TAS during cruise and descent, see below. Classification: Internal Page 53 of 74

54 TAS (kt) Flown Distance (NM) lbs lbs lbs lbs lbs lbs Figure 28: TAS of flights with different mass Figure 28 shows the TAS managed by simulated B737 aircraft flying at same altitude and different masses. It shows that the TAS changed a bit with different mass of flights. However the delta is as low as we consider it as the TAS does not change. The reference mass taken for the normalization is the reference mass given by the BADA files. Below the mass normalization for four simulated B737 flights with different mass is presented. All flights have been simulated with the same cruise altitude and other parameters. The reference mass from BADA for this aircraft type is 55 tons. The following charts show the altitude and the fuel consumption over the flown distance for a flight and its normalization. Important to notice is that the flying time for the two different flights is different, thus the fuel burn for the one that descend earlier is higher than the other, even though both flights is flying on idle, making the fuel consumption to merge closer to touchdown. Classification: Internal Page 54 of 74

55 x 10 4 Weight -> Ref: 55 tons // Real: 45 tons Real ALt (ft) Norm Alt (ft) TOD point Real Fuel (kg) Norm Fuel (kg) Figure 29: Mass normalization for the 45 tons flight 5 x 104 Weight -> Ref: 55 tons // Real: 50 tons Real ALt (ft) Norm Alt (ft) TOD point Real Fuel (kg) Norm Fuel (kg) Figure 30: Mass normalization for the 50 tons flight Classification: Internal Page 55 of 74

56 5 x 104 Weight -> Ref: 55 tons // Real: 59 tons Real ALt (ft) Norm Alt (ft) TOD point Real Fuel (kg) Norm Fuel (kg) Figure 31: Mass normalization for the 59 tons flight 5 x 104 Weight -> Ref: 55 tons // Real: 63 tons Real ALt (ft) Norm Alt (ft) TOD point Real Fuel (kg) Norm Fuel (kg) Figure 32: Mass normalization for the 63 tons flight Classification: Internal Page 56 of 74

57 Below the real and the normalized fuel burns for these simulated flights are presented: 45 tons 50 tons 59 tons 63 tons Real Fuel Burn (kg) Norm Fuel Burn (kg) Delta (%) Standard Mean Deviation Real Norm Classification: Internal Page 57 of 74

58 F. Altitude Normalization Two fligths with the same conditions except for the altitudes have different TAS if operated with the same CI. The aircraft operating at a higher altitude flies faster and has better performance compared to the lower one. However, in order to simplify the normalization, we assume that the performance and thus the fuel flow stay the same for the different altitudes. The normalization focuses on the different location of ToD. An aircraft, operating at a higher altitude (F1 on the Figure 33) starts its descent sooner compared to a lower altitude operating aircraft (F2 on the Figure 33). It will consume less fuel until it reaches the altitude for the lower one. Intermediate Phase F1 F2 Figure 33: Flights flying on different altitude The altitude normalization process only involves the intermediate phase as shown on the Figure 33. We assume that the different flights have the same fuel consumption before and after this phase. The software considers two cases: - the reference altitude is higher than the cruise altitude of the flight - the reference altitude is lower than the cruise altitude of the flight The two following examples below have been done with two simulations flown with the B737 simulator. The first simulation flew at FL370 and the other one flew at FL290 and for these normalizations, the reference altitude was set to FL350. Classification: Internal Page 58 of 74

59 5 x 104 Altitude -> Ref: FL350 // Real: FL370 Real ALt (ft) Norm Alt (ft) TOD point Real Fuel (kg) Norm Fuel (kg) Figure 34: Altitude normalization Simulation flying on FL370 4 x 104 Altitude -> Ref: FL350 // Real: FL290 Real ALt (ft) Norm Alt (ft) TOD point Real Fuel (kg) Norm Fuel (kg) Figure 35: Altitude normalization Simulation flying on FL290 Classification: Internal Page 59 of 74

60 Below is the resulting data for these two simulations: Real Fuel Burn (kg) Normalized Fuel Burn (kg) Delta (kg) FL370 Sim FL290 Sim Classification: Internal Page 60 of 74

61 Annex C. Dissemination event invitation and agenda Classification: Internal Page 61 of 74

62 Classification: Internal Page 62 of 74

63 Classification: Internal Page 63 of 74

64 Annex D. Press Releases The following Press Releases have been published. Classification: Internal Page 64 of 74

65 Classification: Internal Page 65 of 74

66 Classification: Internal Page 66 of 74

67 Classification: Internal Page 67 of 74

68 Classification: Internal Page 68 of 74

69 Classification: Internal Page 69 of 74