Optimisation and Prioritisation of Flows of Air Traffic through an - - PowerPoint PPT Presentation

▶
optimisation and prioritisation of
SMART_READER_LITE
LIVE PREVIEW

Optimisation and Prioritisation of Flows of Air Traffic through an - - PowerPoint PPT Presentation

Dec 26, 2022 645 likes •887 views

Optimisation and Prioritisation of Flows of Air Traffic through an ATM Network SESAR Innovation Days, Braunschweig, November 2012 Hugo de Jonge and Ron Selje, NLR jongehw@nlr.nl Nationaal Lucht- en Ruimtevaartlaboratorium National

slide-1
SLIDE 1

Nationaal Lucht- en Ruimtevaartlaboratorium – National Aerospace Laboratory NLR

Optimisation and Prioritisation of Flows of Air Traffic through an ATM Network

SESAR Innovation Days, Braunschweig, November 2012 Hugo de Jonge and Ron Seljée, NLR jongehw@nlr.nl

slide-2
SLIDE 2

Overview

 The ATM Network in Europe  To enable Optimizing and Prioritizing ATFM by

OPT-ATFM:

 Local context in space and time  How to apply enhanced ATFM

 Model-based experiment:

 Scenario  Slot selection, using prioritization  Prioritization applied on 6 airport flows (5 disrupted)  Results

 Conclusions

2

slide-3
SLIDE 3

Aim to develop prototype implementation of enhanced ATFM algorithm

 Problems in the European ATM Network:

 Systematic overload and congestion of small number of major

airports

 Airports with unbalanced operational conditions, e.g. by weather  Some airports suffering access restriction by lack of capacity of

TMAs, e.g. London area

 Some ATS routes are critical due to En-route constraints, e.g. in

the Core Area

 EC statement:

 “...the use of transparent and efficient rules will provide a flexible and

timely management of air traffic flows at European level and will optimize the use of air routes.”  Objective of this research:

 To demonstrate benefits by regulating departure flows with

enhanced ATFM algorithm

 Aim to apply ATFM with maximum throughput, best achievable

efficiency and minimum impact on flight performance

3

slide-4
SLIDE 4

4

The ATM Network and the Bottlenecks

Network defined by:

  • non-optimal number of

sectors

  • non-optimal distribution
  • f air traffic

Problems:

  • Specific sector nodes

with overload problems

  • Very thick and very thin

traffic flows

  • Airports nodes

very small, very large and sometimes very congested Conclusion:

  • Network is super critical
  • Solving bottlenecks

to make Network robust

  • More robust network

by aggregation

slide-5
SLIDE 5

Enhanced ATFM in adherence to SESAR

 ATFM today:

 ICAO flightplans, possibly updated  Assign overloaded nodes to be regulated  FC-FS: Assign departure slots on overloads by first arrival at the

  • verloaded node (sector)

 Options for Optimization and Prioritization:

 Economic value prevails over FC-FS principle  4D RBTs, up-to-date by SWIM (SESAR)  Apply optimization towards minimized imposed delays  Impose pre-departure delay, whilst being able to assess impact

  • n other flight operations

 Apply optimization towards the economic value of flight by

prioritization

 Alternative solutions:

 MILP and find optimum against minimizing cost-function  Use Petri-Nets and select a local context in space and time

5

slide-6
SLIDE 6

6

Optimising and Prioritising ATFM in Local context of space and time

Proposed solution for improvement: Imposed pre-departure delays Weighted minimisation

  • f imposed delays over 1 sector during 1 hour

1 hour

Time

slide-7
SLIDE 7

Demand and Capacity Balancing (DCB) Solution Strategy

 Full ATM network, including Airports

 Sensitive for unbalance, congestion and overload, but  More accurate and stable load, and thus higher declared capacity

 Local context in space (node) and time (1 hour):

 FC-FS  Pre-departure delay for first flight to hit overloaded node  Select the best one: natural trade-off within local context  Option 1: to select within node with minimized penalties  Option 2: to look at impact of pre-departure delay on other nodes

(before and behind overloaded node)

 Option 3: to give priority to congested flows  Option 4: to give priority to flights on economic value

 Success criteria:

 DCB model (OPT-ATFM): Evaluate minimum pre-departure delay

and minimum “waiting time”

 Operational validation by Fast Time Simulation: Evaluate

minimum flight delay, flight efficiency, workload

7

slide-8
SLIDE 8

8

ATM: DCB Network and Flight operations Network

DCB Network:

  • Airport node capacity:

Sustainable/Max. capacity in mov.

  • Sector node capacity:

Declared capacity

  • Air traffic demand:

Demand per node per hour Flight operations:

  • Airport load:

departure/arrival throughput

  • Sector load:

workload and complexity

  • Air traffic operations:

actual departures

slide-9
SLIDE 9

9

Model-based Experiment on Kernel Network Scenario with 15 main (hub) airports

Main (Hub) Airports Aggregation of smaller airports per country Out Nodes, feeding from outside and functioning as exit nodes for

  • utbounds

Kernel Network defined by a wider area

  • f Europe around the

Core Area. Capacity:

  • 15 main (hub) airports
  • 514 other airports
  • 736 sectors

Demand:

  • 24.600 flights
slide-10
SLIDE 10

OPT-ATFM: slot selection in congested node with minimized imposed delay

10

Example in busy Brussels sector, start of the day, optimized ATFM:

  • In-flight (lila), in-flight via out-node (orange),
  • Flow managed (blue), Flow-managed with penalty in other sector (red)
slide-11
SLIDE 11

Experiment to optimize throughput with 5 disrupted /6 prioritized airports

 Sensitivity analysis of ATM Network throughput:

 Kernel Network, 24.600 flights through Core Area

 Three runs (OPT-ATFM):

– Normal scenario, calculate imposed pre-departure delay – Disrupted scenario, 5 airports disrupted (assumed lower capacity: EHAM, EDDF, EDDM, EGKK, LFPG), calculate imposed pre-departure delay – Disrupted scenario:

 5 airports disrupted (capacity -20% to -30%)  6 airports prioritized (5 disrupted + EGLL added)  Calculate imposed pre-departure delay

 Compare:

– Distribution of “Waiting time” over the day – Geographical distribution of imposed pre-departure delay – Comparing imposed pre-departure delays of large airports

11

slide-12
SLIDE 12

12

Hourly distribution “Waiting time” and Hourly distribution imposed pre-departure delay

100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 5h - 6h 6h - 7h 7h - 8h 8h - 9h 9h - 10h 10h - 11h 11h - 12h 12h - 13h 13h - 14h 14h - 15h 15h - 16h 16h - 17h 17h - 18h 18h - 19h 19h - 20h 20h - 21h 21h - 22h

Hourly Waiting Time (hrs)

Base OPT-ATFCM

total

100 200 300 400 500 600 700 800 900 1000 1100 1200 5h - 6h 6h - 7h 7h - 8h 8h - 9h 9h - 10h 10h - 11h 11h - 12h 12h - 13h 13h - 14h 14h - 15h 15h - 16h 16h - 17h 17h - 18h 18h - 19h 19h - 20h 20h - 21h 21h - 22h 22h - 23h 23h - 24h

Hourly Imposed Predeparture Delays (hrs)

OPT-ATFCM

airports

Example of applying OPT-ATFM with prioritisation:

1.

Measure “Waiting time” over Kernel Network (5 airports disrupted (Blue line left)

2.

Calculate pre-departure delays for 6-airports prioritised scenario (blue line right)

3.

Measure “Waiting time” per node accepting imposed delays (red line left)

slide-13
SLIDE 13

13

5-airports disrupted scenario Apply OPT-ATFM without prioritisation

Kernel Network:

  • Apply OPT-ATFM

ATFM, no prioritisation:

  • #flights “waiting” time:

– ~4.800

  • Total “waiting” time:

– ~12.000 hrs.

  • #flights imposed delay:

– ~5.300

  • Av. imposed delay:

– 34,9 min.

  • At main airports:

– 53,5 min. Conclusion: Waiting time resulting from reduction in capacity at main airports solved mainly by assigning pre- departure delays to these airports.

slide-14
SLIDE 14

14

5-airports disrupted scenario Apply OPT-ATFM with prioritisation (6 airports)

Kernel Network:

  • Apply OPT-ATFM (+ prio)

ATFM, with prioritisation:

  • #flights “waiting” time:

– ~6.000

  • Total “waiting” time:

– ~11.200 hrs.

  • #flights imposed delay:

– ~6.400

  • Av. imposed delay:

– 31,7 min.

  • At main airports:

– 29,9 min. Conclusion: Waiting time resulting from reduction in capacity at main airports distributed and largely assigned to less congested small airports.

slide-15
SLIDE 15

15

Compare OPT-ATFM without and with prioritisation (per airport)

200 400 600 800 1000 1200 1400 1600 1800 2000 2200

Imposed pre-departure delay at 20 most affected airports compared to PrioCase after OPT-ATFCM (hrs)

RefCase ReduMultipleCase PrioCase

Blue: Reference case, no excessive delays Red: No prioritisation, 5-airports disrupted, 5 airports heavy delayed Green: With prioritisation, 6 airports enhanced performance, other airports decreased performance

slide-16
SLIDE 16

16

Some conclusions

 Feasibility evaluated:

 Analysis of ATM network as a network, not by assessment of

  • perational performance

 Assessment of DCB, using 4D trajectory predictions  Flow management within local context of space and time  And .... with optimisation and prioritisation

 Transparency:

 Penalties/benefits per node per priority class

 Experimental results (of first experiment):

 Comparing a 5-airports disrupted scenario with and without

prioritisation to/from congested destinations: – Decrease of total amount of delay – Decrease of average delay – Strong decrease of delay at main airports (-40%)

 All together a more efficient re-distribution of imposed delays

slide-17
SLIDE 17

17

Thank you

Recent reports:

  • NLR-TP-2007-650, Jonge, H.W.G. de, Beers, J.N.P., Seljée, R.R., “Flow Management on

the ATM Network in Europe”, October 2007.

  • NLR-CR-2011-379, Jonge, H.W.G. de, Seljée R.R., “Preliminary validation of Network

Analysis Model (NAM) and Optimising Air Traffic Flow Management (OPT-ATFM)”, December 2011.

  • NLR-TP-2011-567, Jonge, H.W.G. de, Seljée R.R., “Optimisation and Prioritisation of

Flows of Air Traffic through an ATM Network”, December 2011.

slide-18
SLIDE 18

Demand and Capacity Balancing (DCB) Solution Strategy

 Balancing Demand and Capacity (static picture)

 Capacity per sector:

– Over-dimensioned network: Aggregation

 Capacity per Airport:

– Peak capacity, operational achievable throughput – Varying demand/capacity, solved e.g. by hourly capacity

  • ver the day

– No differentiation between departure and arrival flows

 Demand (converging and layered planning):

– 4D Reference Business Trajectories (RBTs) and SWIM  DCB (dynamic picture):

 Flows through network (unconstrained)  Bottlenecks  Flows through network (constrained)  Throughput and “waiting

Time”

 Pre-departure imposed delays  to suppress waiting time  Question: How to optimize “Pre-departure delay” versus

“waiting time”?

18

slide-19
SLIDE 19

19

Network Analysis Model (NAM) Bottleneck analysis

Throughput analysis through ATM networks:

Applicable to sub-networks

Airports and Sectors treated as identical nodes in the network

Used:

To detect hot-spots at airspace and airport level

To analyse balance in the network

To optimise towards minimal “waiting periods”

slide-20
SLIDE 20

Iteration of ATFM: Reduced waiting time by Imposed pre-departure delays

 Conclusion:

Under nominal conditions (no disruption and no connectivity) fairly equalised balance between waiting periods at sectors and at airports (40%- 60%).

20

500 1000 1500 2000 2500 3000 3500 4000 Base FirstFM SecondFM

1368 305 187 2465 691 297

Airports vs Sectors

Number of flights with a waiting time in period at airports Number of flights with a waiting time in period at sectors

slide-21
SLIDE 21

21

Experiment, comparing FC-FS ATFM with OPT-ATFM (PRIO)

Conduct of experiment:

  • 1. Measure “Waiting time” per node
  • 2. Calculate pre-departure delays to mitigate “waiting time”
  • 3. Measure “Waiting time” per node accepting imposed

delays

10 20 30 40 50 60 70 80 90 100 110 120 0h-1h 1h-2h 2h-3h 3h-4h 4h-5h 5h-6h 6h-7h 7h-8h 8h-9h 9h-10h 10h-11h 11h-12h 12h-13h 13h-14h 14h-15h 15h-16h 16h-17h 17h-18h 18h-19h 19h-20h 20h-21h 21h-22h 22h-23h 23h-24h Number of flights per hour

Hourly traffic vs peak capacity

after 30% decrease of capacity

EHAM peak capacity

Example (EHAM) of disruption by (30%) reduction of capacity

slide-22
SLIDE 22

22

Compare ATFM without and with prioritisation

Throughput analysis by measuring “Waiting time”

Total number of flights ATFM FC-FS PrioCase Percent. Number of flights with a waiting time in period 4774 5998 26% Total waiting time ATFM FC-FS PrioCase Percent. Total waiting time in period (hrs) 12003 11246

  • 6%

Calculated imposed pre-departure constraints

Total number of flights ATFM FC-FS PrioCase Percent. Number of flights with a pre-departure delay in period at airports 5324 6387 20% Number of flights with a pre-departure delay in period at main airports 3448 3368

  • 2%

Total pre-departure delay after each run ATFM FC-FS PrioCase Percent. Total pre-departure delay in period at airports (hrs) 12199 11076

  • 9%

Pre-departure delay in period at main airports (hrs) 10010 5589

  • 44%

Average per flight after each run ATFM FC-FS PrioCase Percent. Pre-departure delay in period (min) 34.9 31.7

  • 9%

Pre-departure in period at main airports (min) 53.5 29.9

  • 44%

Pre-departure in period at remaining airports (min) 13.7 34.4 151%

slide-23
SLIDE 23

23