integrated optimization of terminal maneuvering area and airport€¦ · air traffic forecast...

Post on 23-May-2020

2 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Integrated Optimization of Terminal ManeuveringArea and Airport

Ji MA, Daniel DELAHAYE, Mohammed SBIHI, Marcel MONGEAU

ENAC – École Nationale de l’Aviation Civile

SESAR Innovation Days, November 10, 2016

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 1 / 33

Outline

1 Background and problem description

2 Problem modeling

3 Solution approaches

4 Simulation results

5 Conclusions and perspectives

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 2 / 33

Outline

1 Background and problem description

2 Problem modeling

3 Solution approaches

4 Simulation results

5 Conclusions and perspectives

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 3 / 33

Air traffic forecast

According to Airbus global

market forecast 2015-2034,

air traffic will double in the

next 15 years.

39 out of the 47 aviation mega cities are

largely congested today.

airport infrastructure is adequate

airports with potential for congestion

airports where conditions make it

impossible to meet demand

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 4 / 33

Background

Airport and surrounding terminal airspaces control is a complex problem.

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 5 / 33

Current research

Arrival Management Problem

Landing sequencing

Ensure proper separation

Surface Management Problem

Arriving aircraft taxi-in

Departing aircraft taxi-out

Departure Management Problem

Take-off times and sequences for departingflights

Ensure proper separation

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 6 / 33

Outline

1 Background and problem description

2 Problem modeling

3 Solution approaches

4 Simulation results

5 Conclusions and perspectives

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 7 / 33

Given data (1/3)

TMA arrival route networkNode-link graph : G(N ,L ), N node set and L link set ;

One route is composed of several links, the first one starts from the enteringpoint and the last link ends at the runway threshold.

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 8 / 33

Given data (2/3)

Network abstractionOverall terminal capacity : number of gates

Taxi network capacity : threshold of total allowed number of taxi-in and taxi-out aircraft

Runway type : landing only, departure only, mixed mode

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 9 / 33

Given data (3/3)

Given a set of flights, F ={1, . . . ,N f

}. Each flight can be in one of three

operations : F = A∪D ∪AD, where A stands for arrival, D for departure andAD for arrival-departure.

Table : Given information for each operation type

Operation type 00A00 00D00 00AD00

Wake turbulence category√ √ √

Assigned terminal number√ √ √

Entering waypoint√

×√

Initial entry time at TMA√

×√

Initial speed at TMA√

×√

Taxi-in duration√

×√

Earliest off-block time ×√ √

Taxi-out duration ×√ √

Departure runway number ×√ √

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 10 / 33

Decision variables

t f ∈ T f entering time at TMA of flight f ∈ A⋃AD, where

T f = {T 0f + j∆T |∆Tmin/∆T 6 j 6 ∆Tmax/∆T, j ∈ Z}

v f ∈ V f entering speed at TMA of flight f ∈ A⋃AD, where

V f = {Vminf + j∆v

f | j 6 (Vmaxf − Vmin

f )/∆vf , j ∈ N, }

raf ∈ R f landing runway of flight f ∈ A

⋃AD, where

R f = Set of landing runways

p f ∈ P f pushback delay of flight f ∈ D⋃AD, where

P f = {P0f + j∆T |0 6 j 6 ∆T p

max/∆T, j ∈ N}

Decision vector :x = (t, v, r,p)

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 11 / 33

Decision variables

t f ∈ T f entering time at TMA of flight f ∈ A⋃AD, where

T f = {T 0f + j∆T |∆Tmin/∆T 6 j 6 ∆Tmax/∆T, j ∈ Z}

v f ∈ V f entering speed at TMA of flight f ∈ A⋃AD, where

V f = {Vminf + j∆v

f | j 6 (Vmaxf − Vmin

f )/∆vf , j ∈ N, }

raf ∈ R f landing runway of flight f ∈ A

⋃AD, where

R f = Set of landing runways

p f ∈ P f pushback delay of flight f ∈ D⋃AD, where

P f = {P0f + j∆T |0 6 j 6 ∆T p

max/∆T, j ∈ N}

Decision vector :x = (t, v, r,p)

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 11 / 33

Decision variables

t f ∈ T f entering time at TMA of flight f ∈ A⋃AD, where

T f = {T 0f + j∆T |∆Tmin/∆T 6 j 6 ∆Tmax/∆T, j ∈ Z}

v f ∈ V f entering speed at TMA of flight f ∈ A⋃AD, where

V f = {Vminf + j∆v

f | j 6 (Vmaxf − Vmin

f )/∆vf , j ∈ N, }

raf ∈ R f landing runway of flight f ∈ A

⋃AD, where

R f = Set of landing runways

p f ∈ P f pushback delay of flight f ∈ D⋃AD, where

P f = {P0f + j∆T |0 6 j 6 ∆T p

max/∆T, j ∈ N}

Decision vector :x = (t, v, r,p)

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 11 / 33

Separation requirements

Minimum horizontal separation of 3 NM in TMA

Wake turbulence separation

CategoryLeading Aircraft

Heavy Medium Light

Trailing Aircraft

Heavy 4 3 3

Medium 5 3 3

Light 6 5 3

Table : Separation minima for two successive aircraft, in NM

Single-runway separation requirements

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 12 / 33

Conflicts detection

Wake turbulence separation :

Link conflict

Node u Node vLink l = (u,v)

Flight f

Flight f Flight g

Flight g

Minimum horizontal separation :

Node conflict

Node n

Rn

Detection zone

Flight f

Flight g

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 13 / 33

Runway overload evaluation

We note the accumulated time of separation violation for all pairs of aircraft as anindicator for our runway evaluation.

Runway

Heavy Medium Light

Violation of separation

Required time separation

Landing minimum separation times (in seconds)

Pred.\Succ. Heavy Medium Light

Heavy 96 157 207Medium 60 69 123Light 60 69 82

Take-off minimum separation times (in seconds)

Pred.\Succ. Heavy Medium Light

Heavy 96 111 120Medium 60 60 60Light 60 60 60

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 14 / 33

Terminal and taxi network overload evaluationWe measure the maximum overload number and the total amount of time duringwhich aircraft experience congestions.

Num

ber

of

airc

raft

in t

erm

inal

TimeF1

F2

F3

F4

F5

: Aircraft off−block time: Aicraft in−block time

...10:00

1

2

3

4

5

11:0010:10 10:28 10:52 11:2011:15 11:24 11:30 11:50

Capacity=3

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 15 / 33

Objective function

We minimize

γaA(x) + γsS (x)

where

A(x) : the total number of conflicts in airspace, including :

Node conflicts

Link conflicts

S (x) : the airside capacity overload, including :

Runway overload

Terminal overload

Taxi network overload

Weighting coefficients γa, γs

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 16 / 33

Outline

1 Background and problem description

2 Problem modeling

3 Solution approaches

4 Simulation results

5 Conclusions and perspectives

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 17 / 33

Solution approaches

Using time decomposition approach combined with heuristic algorithm.

0

...

Current time window

Previous time window

Next time window

plannedplannedactivecompleted on−going

Update flights status in state space

Evaluate "Active" and "On−going" flight operations:

Airport: runways, taxi network, terminalsAirspace: nodes, links

Apply algorithms:

Simulated Annealing

24 hours

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 18 / 33

Simulated annealing

NUMBER OF ITERATIONS

OB

JEC

TIV

E F

UN

CT

ION

AT INIT TEMP. Unconditional Acceptance

AT FINAL TEMP

HILL CLIMBING

HILL CLIMBING

HILL CLIMBING

Moved accepted with

probability e−∆ET

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 19 / 33

Neighborhood selection (1/3)

Decision Changes

Airspace perfo

Ground perfo

Runway perfo

Aircraft list

A1 Ai AN

R1 Ri RN

G1 Gi GN

xi xNx1

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 20 / 33

Neighborhood selection (2/3)Example (Ground perfo) :

��������������������������������

Nu

mb

er o

f ai

rcra

ft i

n t

erm

inal

TimeF1

F2

F3

F4

F5

: Aircraft off−block time: Aicraft in−block time

...10:00

1

2

3

4

5

11:0010:10 10:28 10:52 11:2011:15 11:24 11:30 11:50

Capacity=3

Flight F1 F2 F3 F4 F5

Ground perfo 32

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 21 / 33

Neighborhood selection (2/3)Example (Ground perfo) :

���������������������������������

���������������������������������

Nu

mb

er o

f ai

rcra

ft i

n t

erm

inal

TimeF1

F2

F3

F4

F5

: Aircraft off−block time: Aicraft in−block time

...10:00

1

2

3

4

5

11:0010:10 10:28 10:52 11:2011:15 11:24 11:30 11:50

Capacity=3

Flight F1 F2 F3 F4 F5

Ground perfo 32 47

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 21 / 33

Neighborhood selection (2/3)Example (Ground perfo) :

�������������������������������������������������������

�������������������������������������������������������

Nu

mb

er o

f ai

rcra

ft i

n t

erm

inal

TimeF1

F2

F3

F4

F5

: Aircraft off−block time: Aicraft in−block time

...10:00

1

2

3

4

5

11:0010:10 10:28 10:52 11:2011:15 11:24 11:30 11:50

Capacity=3

Flight F1 F2 F3 F4 F5

Ground perfo 32 47 47

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 21 / 33

Neighborhood selection (2/3)Example (Ground perfo) :

�����������������������������������

�����������������������������������

Nu

mb

er o

f ai

rcra

ft i

n t

erm

inal

TimeF1

F2

F3

F4

F5

: Aircraft off−block time: Aicraft in−block time

...10:00

1

2

3

4

5

11:0010:10 10:28 10:52 11:2011:15 11:24 11:30 11:50

Capacity=3

Flight F1 F2 F3 F4 F5

Ground perfo 32 47 47 23

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 21 / 33

Neighborhood selection (2/3)Example (Ground perfo) :

Nu

mb

er o

f ai

rcra

ft i

n t

erm

inal

TimeF1

F2

F3

F4

F5

: Aircraft off−block time: Aicraft in−block time

...10:00

1

2

3

4

5

11:0010:10 10:28 10:52 11:2011:15 11:24 11:30 11:50

Capacity=3

Flight F1 F2 F3 F4 F5

Ground perfo 32 47 47 23 0

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 21 / 33

Neighborhood selection (3/3)

NeighborhoodExample :

Flight F1 F2 F3 F4 F5Ground perfo 32 47 47 23 0Percentage 21.5% 31.5% 31.5% 15.4% 0

31.5%

F3

F4

15.5%F1

21.5%

F2

31.5%

Roulette wheel selection

change v f or t f if flight

f ∈ A

change p f if flight f ∈ D

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 22 / 33

Outline

1 Background and problem description

2 Problem modeling

3 Solution approaches

4 Simulation results

5 Conclusions and perspectives

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 23 / 33

Case study

February 7, 2016 : 562 departures, 554 arrivals.

Paris CDG airport layout27R

27L

26R

26L

Terminal 2

Terminal 3

Terminal 1

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 24 / 33

Conflicts evaluation

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 25 / 33

Initial gate occupancy for each terminal

Terminal 1 Terminal 2 Terminal 3

Max gate occupancy 13 95 59

Imposed max capacity 10 90 56

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 26 / 33

Initial gate occupancy for each terminal

Terminal 1 Terminal 2 Terminal 3

Max gate occupancy 13 95 59Imposed max capacity 10 90 56

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 26 / 33

Terminal overload evaluation (1/2)

Figure : Comparison between initial gate occupancy and optimized one for terminal 2Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 27 / 33

Terminal overload evaluation (2/2)

(a) Terminal 1

(b) Terminal 3

Figure : Comparison between initial gate occupancy and optimized one for terminal 1 and 3

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 28 / 33

Taxi network overload evaluation

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 29 / 33

Outline

1 Background and problem description

2 Problem modeling

3 Solution approaches

4 Simulation results

5 Conclusions and perspectives

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 30 / 33

Conclusions

A TMA route network structure and an abstraction model of airportcomponents

Time sliding-window approach combined with simulated annealing

Reaching conflict-free solutions and mitigating the airport overload bytime-slot and speed change

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 31 / 33

Perspectives

Microscopic level optimizationTaxi In and Taxi Out routes

Pushback Time update

This process is repeated by the mean of a sliding time window

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 32 / 33

Thank you for your attention !

Ma, Delahaye, Sbihi, Mongeau (ENAC) Integrated Optimization of TMA and Airport SID Delft 2016 33 / 33

top related