Whither dynamic congestion management ?

Whither

Whither

dynamic

dynamic

congestion management ?

congestion management ?

IREP Symposium 2004

IREP Symposium 2004

Cortina

Cortina

d

d

’

’

Ampezzo

Ampezzo

(

(

Italy

Italy

)

)

Louis WEHENKEL

Louis WEHENKEL

–

–

Universit

Universit

é

é

de Li

de Li

è

è

ge

ge

Outline

Outline

The proposal

The proposal

Static congestion management

Static congestion management

Æ

Æ

Dynamic congestion management

Dynamic congestion management

Some arguments in favor of Dynamic CM

Some arguments in favor of Dynamic CM

ƒ

ƒ

Immediate vs. delayed actions

Immediate vs. delayed actions

ƒ

ƒ

Multiple solutions of SCOPF based static CM

Multiple solutions of SCOPF based static CM

ƒ

ƒ

Market power, gaming

Market power, gaming

ƒ

ƒ

Sub

Sub

-

-

optimality of static CM

optimality of static CM

Feasibility and needs for research

Congestion management

Congestion management

¾

¾

Real-

Real

-time decision making by system

time decision making by system

operators to

operators to

z

z

maintain system security above minimum

maintain system security above minimum

z

z

minimize economic impact

minimize economic impact

z

z

avoid discrimination and market bias

avoid discrimination and market bias

¾

¾

Remarks

Remarks

z

z

time step typically of 15 to 30 minutes

time step typically of 15 to 30 minutes

z

z

objectives monitored over a longer term

objectives monitored over a longer term

Static CM

Static CM

¾

¾

Problem formulation

Problem formulation

At time k, compute decision

At time k, compute decision

u

u

_kk

to

to

z

z satisfy satisfy instantaneousinstantaneoussecurity constraints and security constraints and z

z minimize minimize instantaneousinstantaneouscostscosts

Given

Given

x

x

_kk

, select

, select

u

u

_kk

in

in

U(x

U(x

_kk

) to minimize

) to minimize

c

c

_kk

(x

(x

_kk

,

,

u

u

_kk

)

)

¾

¾

Notation

Notation

k : discrete time index (say 15 minutes time step)

k : discrete time index (say 15 minutes time step)

x

x_kk: electrical state vector at time k (estimated): electrical state vector at time k (estimated)

u

u_kk: decision variables at time k (to determine): decision variables at time k (to determine)

c

c_kk(x,u(x,u) : instantaneous cost function, ) : instantaneous cost function, U(x

U(x) : constraints on decision variables ) : constraints on decision variables

¾

¾

NB:

NB:

Constraints Constraints U(xU(x) include feasibility and all kinds of security ) include feasibility and all kinds of security

constraints deemed relevant (N, N

Dynamic CM

Dynamic CM

¾

¾

Problem formulation

Problem formulation

z

z

(Assuming a state transition model :

(Assuming a state transition model : x

x

_{k+1 k+1}

=

=

f(x

f(x

_kk

,

,

u

u

_kk

,

,

_{, ,}

w

w

_{k k}

)

)

)

)

z

z

At time k, compute a

At time k, compute a

sequence

sequence

of decisions

of decisions

(

(

u

u

_kk

, u

, u

_k+1k+1

,

,

…

…

, u

, u

_k+Tk+T--11

)

)

to

to

•

•

satisfy security constraints and

satisfy security constraints and

•

•

minimize cost induced

minimize cost induced

over the whole time interval

over the whole time interval

(k, k+T

(k, k+T

-

-

1)

1)

¾

¾

Remarks

Remarks

z

z

Apply

Apply

u

u

_kk

and refresh the sequence at the next time step

and refresh the sequence at the next time step

z

z

Time step of 5 to 30 minutes

Time step of 5 to 30 minutes

z

z

Time horizon T of several hours to a few days

Time horizon T of several hours to a few days

z

z

(and we neglect faster dynamics)

(and we neglect faster dynamics)

Argument/Question No1

Argument/Question No1

¾

¾

Dynamic Congestion Management can in

Dynamic Congestion Management can in

principle reach a better security/cost

principle reach a better security/cost

tradeoff (over the long term) than static

tradeoff (over the long term) than static

congestion management

congestion management

Would savings be significant ?

Would savings be significant ?

Argument/Question No2

Argument/Question No2

¾

¾

In dynamic CM one can include constraints

In dynamic CM one can include constraints

and/or penalization terms to

and/or penalization terms to

z

z

allocate impact on market participants in a non

allocate impact on market participants in a non

discriminating way (whatever that means

discriminating way (whatever that means

…

…

)

)

z

z

fight against gaming and market power

fight against gaming and market power

What is the impact of this on the

What is the impact of this on the

overall security/cost tradeoff ?

overall security/cost tradeoff ?

Computational feasibility ?

Computational feasibility ?

Try to combine existing model

Try to combine existing model-

-based tools

based tools

z

z

SCOPF, DSA, DSE

SCOPF, DSA, DSE

with stochastic optimization and

with stochastic optimization and

approximation techniques to solve

approximation techniques to solve

stochastic dynamic programming problem

stochastic dynamic programming problem

Can be used on