optimization in modern power systems - chatzivachatziva.com/31xxx/lectures/lecture4.pdf ·...
TRANSCRIPT
![Page 1: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/1.jpg)
Optimization in modern power systems
Lecture 4: Lagrangian and Nodal Prices
Spyros Chatzivasileiadis
Some slides of this lecture have been in-spired or taken from the lecture slides ofGabriela Hug for the class 18-879 M: Op-timization in Energy Networks, CarnegieMellon University, USA, 2015.
![Page 2: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/2.jpg)
Groups and Topics for Assignment 2
1 Primal-dual interior-point method:
2 Simplex method:
3 Newton’s method for optimization with equality constraints:
4 Gradient descent method for unconstrained optimization:
• Peer-review groups
• #1 with #3• #2 with #4
2 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 3: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/3.jpg)
The Goals for Today!
• Review of Day 3
• Questions and Clarifications on Assignments
• Lagrangian for Inequality Constrained Optimization
• Extracting the Lagrangian Multipliers (= nodal prices) for the DC-OPFproblem
3 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 4: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/4.jpg)
Reviewing Day 3 in Groups!
• For 10 minutes discuss with theperson sitting next to you about:
• Three main points we discussedin yesterday’s lecture
• One topic or concept that is notso clear to you and you wouldlike to hear again about it
4 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 5: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/5.jpg)
Points you would like to discuss?
Questions about the Assignments?
5 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 6: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/6.jpg)
Notes on Assigment 1
• The line flow constraints of the DC-OPF must be considered for bothdirections
6 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 7: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/7.jpg)
Convex and Concave Functions
• Convex function: a lineconnecting two points must lieabove the function
x
f(x)
• Concave function: a lineconnecting two points must liebelow the function
x
f(x)
• Ideally, we want to minimize convex functions and maximize concavefunctions
7 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 8: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/8.jpg)
Formulating an optimization problem
Example: James, a CMU student, opens a new sandwich shop on CMUcampus to earn some money. He o↵ers two types of sandwiches, tunaand chicken. His costs for the tuna sandwich are $4, his profit is $3.5and it takes him 8 minutes to make one. The costs for the chickensandwich are $6, his profit is $3 and it takes him 6 minutes to makeone. Besides studying, he is able to spend 3 hours per day preparingsandwiches and he has a budget of $120 per day. The universityregulations say that he has to sell at least 5 sandwiches of each type.
• Assuming that James can sell all his sandwiches, write down theoptimization problem to find the number of sandwiches of each typewhich maximize his profit.
• Answer: 15 tuna sandwiches, 10 chicken sandwiches
8 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 9: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/9.jpg)
Formulating an optimization problem
Example: James, a CMU student, opens a new sandwich shop on CMUcampus to earn some money. He o↵ers two types of sandwiches, tunaand chicken. His costs for the tuna sandwich are $4, his profit is $3.5and it takes him 8 minutes to make one. The costs for the chickensandwich are $6, his profit is $3 and it takes him 6 minutes to makeone. Besides studying, he is able to spend 3 hours per day preparingsandwiches and he has a budget of $120 per day. The universityregulations say that he has to sell at least 5 sandwiches of each type.
• Assuming that James can sell all his sandwiches, write down theoptimization problem to find the number of sandwiches of each typewhich maximize his profit.
• Answer: 15 tuna sandwiches, 10 chicken sandwiches
8 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 10: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/10.jpg)
Formulating an optimization problem
Example: James, a CMU student, opens a new sandwich shop onCMU campus to earn some money. He o↵ers two types of sandwiches,tuna and chicken...
• Assuming that James can sell all his sandwiches, write down theoptimization problem to find the number of sandwiches of each typewhich maximize his profit.
• Answer: 15 tuna sandwiches, 10 chicken sandwiches
• Note that this is normally a mixed integer linear problem (MILP). In ourcase, we relax our problem and assume that the optimization variablesare continuous variables. This allows us to solve it with linprog. Wewere“lucky”and the solver returned integers as the optimal result. If wedid not obtain integers, our solution would have been infeasible for theoriginal problem. Then we would need to use di↵erent methods to solveit, e.g. using the intlinprog from the Matlab Optimization Toolbox.
9 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 11: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/11.jpg)
Equality Constrained Optimization
• Example:
minx
x
21 + x
22
s.t.� x1 � x2 + 4 = 0
• Find the solution to this problem using KKT conditions.
10 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 12: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/12.jpg)
Inequality Constrained Optimization
• Find a solution to:
minx
f0(x)
s.t. f
i
(x) = 0 for i=1,. . . ,m
@f0(x)
@x
6= 0 in feasible region
11 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 13: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/13.jpg)
Binding Constraint
• Constraint binding, i.e. fi
(x⇤) = 0
rf0(x) +mX
i=1
�
i
f
i
(x)
f
i
0 for i = 1, . . . ,m
• gradients of objective function and ofconstraint are in opposite directions inoptimal point
) �
i
> 0
• Sensitivity: �i
= ��f0(x)
�f
i
(x)
12 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 14: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/14.jpg)
Non-binding constraint
• Constraint non-binding, i.e. fi
(x⇤) < 0
rf0(x) +mX
i=1
�
i
f
i
(x)
f
i
0 for i = 1, . . . ,m
• gradient of objective function is zero,i.e. rf(x) = 0
) �
i
= 0
• The Lagrange multiplier is:
•>0, for binding inequality constraints
• =0, for non-binding inequality constraints
13 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 15: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/15.jpg)
KKTs for Inequality Constrained Optimization
• Lagrange function
L = f0(x) +mX
i=1
�
i
f
i
(x)
• Minimize Lagrange function
@L
@x
= 0 ) @f(x)
@x
+mX
i=1
�
i
@f
i
(x)
@x
= 0
@L
@�
i
= 0 ) f
i
(x) 0 for i = 1, . . . ,m
�
i
f
i
(x) = 0 for i+ 1, . . . ,m
�
i
� 0
Solution can befound by checkingcombinations ofbinding and non-binding constraints) use solutionalgorithms
14 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 16: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/16.jpg)
KKTs for Constrained Optimization
• Minimize Lagrange function:
L = f0(x) +mX
i=1
�
i
f
i
(x) +pX
i=1
⌫
i
h
i
(x)
• The Karush-Kuhn-Tacker first order or necessary optimality conditions:
@L
@x
= 0 ) @f(x)
@x
+mX
i=1
�
i
@f
i
(x)
@x
+pX
i=1
⌫
i
@h
i
(x)
@x
= 0
@L
@�
i
= 0 ) f
i
(x) 0 for i = 1, . . . ,m
@L
@⌫
i
= 0 ) h
i
(x) = 0 for i = 1, . . . , p
�
i
f
i
(x) = 0 for i+ 1, . . . ,m
�
i
� 0
15 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 17: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/17.jpg)
Costrained Optimization: Example
minx1,x2
(x1 � 3)2 + (x2 � 2)2
subject to:
2x1 + x2 = 8
x1 + x2 7
x1 � 0.25x22 0
• Write down the KKT conditions for this problem.
16 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 18: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/18.jpg)
Constrained Optimization: Graphical Solution
Example:
minx1,x2
(x1 � 3)2 + (x2 � 2)2
subject to:
2x1 + x2 = 8
x1 + x2 7
x1 � 0.25x22 0
x1 � 0
x2 � 0
17 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 19: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/19.jpg)
Mathematical Formulations: Summary
Unconstrained Optimization minx
f(x)
Equality Constrained Optimization minx
f(x)
h
i
(x) = 0 for i = 1, . . . , p
Inequality Constrained Optimiza-tion
minx
f(x)
f
i
(x) 0 for i = 1, . . . ,m
General Constrained Optimizationminx
f(x)
f
i
(x) 0 for i = 1, . . . , p
h
i
(x) = 0 for i = 1, . . . ,m
18 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 20: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/20.jpg)
Convex Optimization
• The optimization problem
minx
f(x)
f
i
(x) 0 for i = 1, . . . , p
h
i
(x) = 0 for i = 1, . . . ,m
is convex if:
• the objective function f(x) is convex
• the inequality constraints fi
(x) are convex
• the equality constraints hi
(x) are linear
If the problem is convex, there is a single optimum, which is also the
global optimum!
19 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 21: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/21.jpg)
Solution Types for Linear OptimizationUnique Solution Unbounded Solution
Infinitely many solutions No solution
20 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 22: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/22.jpg)
DC-OPF based on PTDF
min
NPGX
i=1
c
i
P
G,i
,
subject to:NPGX
i=1
P
G,i
�NPLX
i=1
P
L,i
= 0
�FL
PTDF · (PG
�PL
) FL
0 PG
PG,max
21 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 23: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/23.jpg)
Lagrangian of the DC-OPF
L(PG
, ⌫,�, µ) =
NPGX
i=1
c
i
P
G,i
+ ⌫ ·
0
@NPGX
i=1
P
G,i
�NPLX
i=1
P
L,i
1
A
+NLX
i=1
�
+i
· [PTDFi
· (PG
�PL
)� F
L,i
]
+NLX
i=1
�
�i
· [�PTDFi
· (PG
�PL
)� F
L,i
]
+
NPGX
i=1
µ
+i
· (PG,i
� P
G,i,max
) +
NPGX
i=1
µ
�i
· (�P
G,i
)
22 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 24: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/24.jpg)
Test System
• Assume a 3-bus system with 3 generators, and 1 load on bus 3
• We assume an auxilliary variable ⇠3 that represents very small changes ofthe load in Bus 3. We assume ⇠3 = 0.
• Then it is P̂L
= P
L
+ ⌅, where ⌅ = [0 0 ⇠3]T .
1 2
3
23 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 25: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/25.jpg)
Lagrangian of the DC-OPF with ⌅
L(PG
, ⌫,�, µ,⌅) =
NPGX
i=1
c
i
P
G,i
+ ⌫ ·
0
@NPGX
i=1
P
G,i
�NPLX
i=1
P
L,i
� ⇠
i
1
A
+NLX
i=1
�
+i
· [PTDFi
· (PG
�PL
� ⌅)� F
L,i
]
+NLX
i=1
�
�i
· [�PTDFi
· (PG
�PL
� ⌅)� F
L,i
]
+
NPGX
i=1
µ
+i
· (PG,i
� P
G,i,max
) +
NPGX
i=1
µ
�i
· (�P
G,i
).
24 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 26: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/26.jpg)
Lagrangian of DC-OPF for the 3-bus system
• To save space in this slide: Ki
⌘ PTDF
i
L(PG
, ⌫,�, µ, ⇠3) =
NPGX
i=1
c
i
P
G,i
+ ⌫ ·
0
@NPGX
i=1
P
G,i
�NPLX
i=1
P
L,i
� ⇠3
1
A
+NLX
i=1
�
+i
· [Ki,1 · PG,1 +K
i,2 · PG,2 +K
i,3 · (PG,3 � P
L,3 � ⇠3)� F
L,i
]
+NLX
i=1
�
�i
· [�K
i,1 · PG,1 �K
i,2 · PG,2 �K
i,3 · (PG,3 � P
L,3 � ⇠3)� F
L,i
]
+
NPGX
i=1
µ
+i
· (PG,i
� P
G,i,max
) +
NPGX
i=1
µ
�i
· (�P
G,i
).
25 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 27: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/27.jpg)
KKTs for the DC-OPF: No congestion
• No congestion ) all �i
= 0
• One marginal generator: only one generator has both µ
+i
= 0 and µ
�i
= 0
• Assume G2 is marginal; PG1 = P
G1,max
; PG3 = 0.
@L@P
G,i
= 0, for all i 2 N
PG
c1 + ⌫ + µ
+1 = 0
c2 + ⌫ = 0
c3 + ⌫ + µ
�3 = 0
Marginal change in the cost func-tion for a marginal change in load:
LMP3 =@L@⇠3
= �⌫
Attention! ⇠3 does not exist in the optimization problem and is not anoptimization variable. We do not need to derive any KKT conditions
w.r.t. ⇠3, e.g.@L
@⇠3= 0.
⇠3 is just an auxilliary variable. It helps us“represent” the marginalchange in the load of bus 3. @L
@⇠3quantifies its e↵ect on the Lagrangian.
26 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 28: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/28.jpg)
KKTs for the DC-OPF: No congestion
• No congestion ) all �i
= 0
• One marginal generator: only one generator has both µ
+i
= 0 and µ
�i
= 0
• Assume G2 is marginal; PG1 = P
G1,max
; PG3 = 0.
@L@P
G,i
= 0, for all i 2 N
PG
c1 + ⌫ + µ
+1 = 0
c2 + ⌫ = 0
c3 + ⌫ + µ
�3 = 0
Marginal change in the cost func-tion for a marginal change in load:
LMP3 =@L@⇠3
= �⌫
Attention! ⇠3 does not exist in the optimization problem and is not anoptimization variable. We do not need to derive any KKT conditions
w.r.t. ⇠3, e.g.@L
@⇠3= 0.
⇠3 is just an auxilliary variable. It helps us“represent” the marginalchange in the load of bus 3. @L
@⇠3quantifies its e↵ect on the Lagrangian.
26 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 29: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/29.jpg)
KKTs for the DC-OPF: No congestion
• No congestion ) all �i
= 0
• One marginal generator: only one generator has both µ
+i
= 0 and µ
�i
= 0
• Assume G2 is marginal; PG1 = P
G1,max
; PG3 = 0.
@L@P
G,i
= 0, for all i 2 N
PG
c1 + ⌫ + µ
+1 = 0
c2 + ⌫ = 0
c3 + ⌫ + µ
�3 = 0
Marginal change in the cost func-tion for a marginal change in load:
LMP3 =@L@⇠3
= �⌫
Attention! ⇠3 does not exist in the optimization problem and is not anoptimization variable. We do not need to derive any KKT conditions
w.r.t. ⇠3, e.g.@L
@⇠3= 0.
⇠3 is just an auxilliary variable. It helps us“represent” the marginalchange in the load of bus 3. @L
@⇠3quantifies its e↵ect on the Lagrangian.
26 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 30: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/30.jpg)
KKTs for the DC-OPF: No congestion
• No congestion ) all �i
= 0
• One marginal generator: only one generator has both µ
+i
= 0 and µ
�i
= 0
• Assume G2 is marginal; PG1 = P
G1,max
; PG3 = 0.
@L@P
G,i
= 0, for all i 2 N
PG
c1 + ⌫ + µ
+1 = 0
c2 + ⌫ = 0
c3 + ⌫ + µ
�3 = 0
Marginal change in the cost func-tion for a marginal change in load:
LMP3 =@L@⇠3
= �⌫
Attention! ⇠3 does not exist in the optimization problem and is not anoptimization variable. We do not need to derive any KKT conditions
w.r.t. ⇠3, e.g.@L
@⇠3= 0.
⇠3 is just an auxilliary variable. It helps us“represent” the marginalchange in the load of bus 3. @L
@⇠3quantifies its e↵ect on the Lagrangian.
26 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 31: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/31.jpg)
KKTs for the DC-OPF: No congestion
• No congestion ) all �i
= 0
• One marginal generator: only one generator has both µ
+i
= 0 and µ
�i
= 0
• Assume G2 is marginal; PG1 = P
G1,max
; PG3 = 0.
@L@P
G,i
= 0, for all i 2 N
PG
c1 + ⌫ + µ
+1 = 0
c2 + ⌫ = 0
c3 + ⌫ + µ
�3 = 0
Marginal change in the cost func-tion for a marginal change in load:
LMP3 =@L@⇠3
= �⌫
LMP3 = �⌫ = c2: nodal price on bus 3!How much is the LMP on the other buses?
27 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 32: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/32.jpg)
KKTs for the DC-OPF: One congested line
• Assume that line 1-3 gets congested in the direction 1 ! 3 ) �
+13 6= 0
• Now G2 and G3 are both marginal gens; PG1 = P
G1,max
.
@L@P
G,i
= 0, for all i 2 N
PG
c1 + ⌫ + µ
+1 + �
+13PTDF13,1 = 0
c2 + ⌫ + �
+13PTDF13,2 = 0
c3 + ⌫ + �
+13PTDF13,3 = 0
Marginal change in the cost func-tion for a marginal change in load:
LMP3 =@L@⇠3
= �⌫��
+13PTDF13,3
To find LMP3 I need ⌫ and �
+13
How do I find ⌫ and �
+13?
28 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 33: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/33.jpg)
KKTs for the DC-OPF: One congested line
• Assume that line 1-3 gets congested in the direction 1 ! 3 ) �
+13 6= 0
• Now G2 and G3 are both marginal gens; PG1 = P
G1,max
.
@L@P
G,i
= 0, for all i 2 N
PG
c1 + ⌫ + µ
+1 + �
+13PTDF13,1 = 0
c2 + ⌫ + �
+13PTDF13,2 = 0
c3 + ⌫ + �
+13PTDF13,3 = 0
Marginal change in the cost func-tion for a marginal change in load:
LMP3 =@L@⇠3
= �⌫��
+13PTDF13,3
To find LMP3 I need ⌫ and �
+13
How do I find ⌫ and �
+13?
28 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 34: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/34.jpg)
KKTs for the DC-OPF: One congested line
• Assume that line 1-3 gets congested in the direction 1 ! 3 ) �
+13 6= 0
• Now G2 and G3 are both marginal gens; PG1 = P
G1,max
.
@L@P
G,i
= 0, for all i 2 N
PG
c1 + ⌫ + µ
+1 + �
+13PTDF13,1 = 0
c2 + ⌫ + �
+13PTDF13,2 = 0
c3 + ⌫ + �
+13PTDF13,3 = 0
Marginal change in the cost func-tion for a marginal change in load:
LMP3 =@L@⇠3
= �⌫��
+13PTDF13,3
To find LMP3 I need ⌫ and �
+13
How do I find ⌫ and �
+13?
28 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 35: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/35.jpg)
KKTs for the DC-OPF: One congested line
• Assume that line 1-3 gets congested in the direction 1 ! 3 ) �
+13 6= 0
• Now G2 and G3 are both marginal gens; PG1 = P
G1,max
.
@L@P
G,i
= 0, for all i 2 N
PG
c1 + ⌫ + µ
+1 + �
+13PTDF13,1 = 0
c2 + ⌫ + �
+13PTDF13,2 = 0
c3 + ⌫ + �
+13PTDF13,3 = 0
Marginal change in the cost func-tion for a marginal change in load:
LMP3 =@L@⇠3
= �⌫��
+13PTDF13,3
To find LMP3 I need ⌫ and �
+13
How do I find ⌫ and �
+13?
28 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 36: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/36.jpg)
KKTs for the DC-OPF: One congested line
• Solve the linear system with 2 equations and 2 unknowns: ⌫ and �
+13
c2 + ⌫ + �
+13PTDF13,2 = 0
c3 + ⌫ + �
+13PTDF13,3 = 0
————————————————————
• What can we say about the LMPs on di↵erent buses?
LMP
i
= �⌫ � �
+13PTDF13,i
• If there is a congestion, the LMPs are no longer the same on every bus.They are dependent on the congestion!
29 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 37: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/37.jpg)
KKTs for the DC-OPF: One congested line
• Solve the linear system with 2 equations and 2 unknowns: ⌫ and �
+13
c2 + ⌫ + �
+13PTDF13,2 = 0
c3 + ⌫ + �
+13PTDF13,3 = 0
————————————————————
• What can we say about the LMPs on di↵erent buses?
LMP
i
= �⌫ � �
+13PTDF13,i
• If there is a congestion, the LMPs are no longer the same on every bus.They are dependent on the congestion!
29 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017
![Page 38: Optimization in modern power systems - Chatzivachatziva.com/31xxx/Lectures/Lecture4.pdf · 2017-03-07 · Optimization in modern power systems Lecture 4: Lagrangian and Nodal Prices](https://reader030.vdocuments.site/reader030/viewer/2022040920/5e96b7fbcc43e754ac071aae/html5/thumbnails/38.jpg)
KKTs for the DC-OPF: One congested line
• Solve the linear system with 2 equations and 2 unknowns: ⌫ and �
+13
c2 + ⌫ + �
+13PTDF13,2 = 0
c3 + ⌫ + �
+13PTDF13,3 = 0
————————————————————
• What can we say about the LMPs on di↵erent buses?
LMP
i
= �⌫ � �
+13PTDF13,i
• If there is a congestion, the LMPs are no longer the same on every bus.They are dependent on the congestion!
29 DTU Electrical Engineering Optimization in modern power systems Jan 5, 2017