29.01.2015 Views

Power System Security Analysis With Renewable Energy ... - IIT Mandi

Power System Security Analysis With Renewable Energy ... - IIT Mandi

Power System Security Analysis With Renewable Energy ... - IIT Mandi

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

<strong>Power</strong> <strong>System</strong> <strong>Security</strong> <strong>Analysis</strong> <strong>With</strong><br />

<strong>Renewable</strong> <strong>Energy</strong> <strong>System</strong>s<br />

Trapti Jain<br />

School of Computing and Electrical Engineering<br />

Indian Institute of Technology <strong>Mandi</strong>


Outline<br />

• <strong>Power</strong> <strong>System</strong> <strong>Security</strong><br />

• <strong>Security</strong> in Restructured <strong>Power</strong> <strong>System</strong><br />

• Review of Load Flow <strong>Analysis</strong><br />

• Load Flow <strong>Analysis</strong> with Wind Farm<br />

• Load Flow <strong>Analysis</strong> with Photovoltaic <strong>System</strong>


<strong>Power</strong> <strong>System</strong> <strong>Security</strong><br />

•It is the ability of the system to withstand credible<br />

contingencies without violating the normal operating<br />

limits.<br />

• Classified as<br />

Static security<br />

static security evaluation detects any potential<br />

overload of a system branch or an out of limit voltage following a<br />

given list of contingencies<br />

Transient security<br />

transient security evaluation pertains to system<br />

dynamic behaviour in terms of rotor angle stability when<br />

subjected to perturbations


<strong>Security</strong> <strong>Analysis</strong><br />

Outages of component(s)<br />

Overstress on the other<br />

components<br />

No limit violation<br />

limit violation(s)<br />

operation of protective<br />

devices and switching of the<br />

unit(s)<br />

partial or total loss of<br />

load


Functions of <strong>System</strong> <strong>Security</strong><br />

<strong>System</strong><br />

Monitoring<br />

<strong>System</strong><br />

<strong>Security</strong><br />

Contingency<br />

<strong>Analysis</strong><br />

<strong>Security</strong><br />

Constrained<br />

OPF


Contingency <strong>Analysis</strong> Procedure<br />

START<br />

SET SYSTEM MODEL TO<br />

INITIAL CONDITIONS<br />

SIMULATE AN OUTAGE OF A<br />

GENERATOR OR A BRANCH<br />

SELECT A<br />

NEW OUTAGE<br />

LIMIT VIOLATION<br />

Y<br />

ALARM MESSAGE<br />

N<br />

N<br />

LAST OUTAGE<br />

END<br />

Y


Real-time applications require fast and reliable computation methods due to the high<br />

number of possible outages in a moderate power system.<br />

However, there is a well-known conflict between the accuracy of the method applied<br />

and the calculation speed.<br />

Exact solution<br />

Full AC power flow<br />

for each outage<br />

Check the limit<br />

violations<br />

not feasible<br />

for realtime<br />

applications<br />

.<br />

approximate methods to quickly<br />

identify conceivable<br />

contingencies<br />

real-time applications<br />

AC power flows only for<br />

critical contingencies.<br />

Check the limit violations


APPROXIMATE CONTINGENCY ANALYSIS<br />

Contingency ranking<br />

• contingencies are ranked in an approximate order<br />

of a scalar performance index, PI.<br />

• contingencies are tested beginning with the most<br />

severe one and proceeding down to the less<br />

severe ones up to a threshold value.<br />

• Masking effect causes false orderings and<br />

misclassifications.<br />

Contingency screening<br />

• Explicit contingency screening is performed for all<br />

contingencies, following an approximate solution<br />

(DC load flow, one iteration load flow, linear<br />

distribution or sensitivity factors etc.)<br />

• Contingency screening is performed in the near<br />

vicinity of the outages (local solutions)<br />

Hybrid methods utilizing both the ranking and the screening


<strong>Security</strong> in Restructured <strong>Power</strong> <strong>System</strong>s<br />

•In deregulated environment, <strong>System</strong> Operator (SO) is<br />

responsible for secure operation of the power system.<br />

•Available Transfer Capability is a term defined to<br />

reflect additional secure power transfer in the electricity<br />

market environment.<br />

•Congestion management is one of the major task<br />

performed by SO to ensure system security.


Available Transfer Capability (ATC) Definition *<br />

“ATC is a measure of the transfer capability remaining in the<br />

physical transmission network for further commercial activity over<br />

and above already committed use”.<br />

Mathematically,<br />

ATC = TTC – TRM – {ETC + CBM}<br />

ETC: Existing Transfer Commitment<br />

TTC: maximum amount of power which can be transferred over the<br />

network while satisfying all security constraints.<br />

TRM: margin required for uncertainties in the transmission system<br />

conditions<br />

CBM: margin reserved by load serving entities for generation<br />

reliability requirements.<br />

*<br />

North American Electric Reliability Council (NERC), Available Transfer Capability Definitions and<br />

Determination, NERC Report, June 1996.


ATC and Related Terms (NERC Report)<br />

TRM – Transfer Reliability Margin<br />

CBM – Capacity Benefit Margin<br />

ATC – Available Transfer Capability<br />

TTC – Total Transfer Capability<br />

ETC - Existing Transfer Commitments<br />

Usually, the TRM and CBM are decided by the utilities according to their<br />

specific system condition and reliability requirements. Thus, the ATC can<br />

be defined as,<br />

ATC = TTC – ETC


ATC Determination<br />

1. Static ATC Methods<br />

• Line flow limits<br />

• Bus voltage limits<br />

•Generator real and reactive power<br />

limits<br />

• Static voltage stability limit<br />

• Repetitive power flow based methods<br />

•Continuation power flow based<br />

methods<br />

• Sensitivity based methods<br />

•Optimal power flow based methods<br />

2. Dynamic ATC Methods<br />

• Line flow limits<br />

• Bus voltage limits<br />

•Generator real and reactive power<br />

limits<br />

• Small signal stability limit<br />

• Large signal stability limit<br />

• Trajectory sensitivity based<br />

methods<br />

• <strong>Energy</strong> function based methods<br />

• Optimal power flow based<br />

methods<br />

• Single Machine Equivalent<br />

(SIME)


Methods of Congestion Management<br />

• Price Area based Congestion Management<br />

- areas with excess generation will have lower prices,<br />

and those with excess load will have higher prices<br />

• ATC based Congestion Management<br />

- use ATC information available there to determine if<br />

system could accommodate transaction.<br />

• OPF based Congestion Management<br />

- optimal power flow (OPF) is performed to minimize<br />

generators’ operating cost subject to set of constraints that<br />

represent a model of the transmission system within which the<br />

generator operate


Objective of Load Flow <strong>Analysis</strong><br />

•To determine the static operating state of the<br />

power system for given loads<br />

•To determine voltage magnitude and angle at<br />

all the buses<br />

•To determine line flows and system losses


Classification of load flow methods<br />

• Deterministic Load Flow (DLF)<br />

- a set of deterministic values are chosen by<br />

the analyst for each input variable<br />

- accuracy depends on the knowledge of the<br />

input data<br />

- ignores grid uncertainty<br />

• Probabilistic Load Flow (PLF)<br />

- utilizes different mathematical approaches<br />

such as probabilistic approach, fuzzy sets, interval<br />

analysis etc. for taking into account uncertainty<br />

- requires inputs with PDF or CDF


DLF with Conventional Generators<br />

Main steps to perform load flow analysis<br />

• Develop the mathematical model of various elements<br />

•Write the network equations relating current and<br />

voltages<br />

•Identify the nature of equations and apply suitable<br />

solution technique


Mathematical Model of Various Elements<br />

• Transmission line - nominal equivalent π model is used<br />

• Transformers - equivalent π model is used<br />

⎧A<br />

=<br />

⎪<br />

⎨B<br />

=<br />

⎪<br />

⎩C<br />

=<br />

y<br />

12<br />

/ a<br />

(1/ a −1)<br />

⋅ y<br />

(1 −1/<br />

a).<br />

y<br />

12<br />

12<br />

/ a<br />

• Loads - constant power model<br />

• Generators - real and reactive power output l


Bus Admittance Matrix or Y bus<br />

• First step in solving the power flow is to create the bus<br />

admittance matrix, often called the Y bus<br />

.<br />

• The Y bus<br />

gives the relationships between all the bus<br />

current injections, I, and all the bus voltages, V,<br />

I = Y bus<br />

V<br />

• The Y bus<br />

is developed by applying KCL at each bus in the<br />

system to relate the bus current injections, the bus<br />

voltages, and the branch impedances and admittances.<br />

• The diagonal elements are the self admittance terms,<br />

equal to the algebraic sum of all the primitive value of<br />

admittances incident at the node or bus.<br />

• The off-diagonal terms are equal to the negative of the<br />

admittance connected between the two buses.


Load Flow Equations<br />

Applying KCL at each bus-i, in an n-bus system, the<br />

current injection I i<br />

, is given by<br />

I<br />

i<br />

=<br />

I<br />

ij<br />

+<br />

I<br />

ik<br />

+ ....... I<br />

in<br />

=<br />

n<br />

∑<br />

v=<br />

1<br />

I<br />

iv<br />

Since I=Y bus<br />

V, we also know<br />

I<br />

i<br />

=<br />

n<br />

∑<br />

j=<br />

1<br />

The net complex power injected into a bus-i is given by<br />

*<br />

S = P + jQ = V I<br />

i<br />

i<br />

i<br />

Taking complex conjugate of above equation,<br />

*<br />

P − jQ = V I<br />

i<br />

i<br />

Y<br />

ij<br />

V<br />

i<br />

i<br />

j<br />

i<br />

i


Load Flow Equations, Contd.<br />

*<br />

( Pi<br />

− jQi<br />

) = ( PGi<br />

− PDi<br />

) − j( QGi<br />

−QDi<br />

) = Vi<br />

YijV<br />

j<br />

In polar form,<br />

( P − jQ ) = ( P − P ) − j( Q − Q )<br />

=<br />

i<br />

n<br />

∑<br />

j=<br />

1<br />

VV Y<br />

i<br />

i<br />

j<br />

ij<br />

Gi<br />

V Y V<br />

Separating into real and imaginary parts<br />

n<br />

∑<br />

j=<br />

1<br />

[ cos( δ −δ<br />

−θ<br />

) − j sin( δ −δ<br />

−θ<br />

)] i = 1,2,.... n<br />

i<br />

Di<br />

j<br />

ij<br />

Gi<br />

Di<br />

=<br />

i<br />

n<br />

∑<br />

j=<br />

1<br />

j<br />

i<br />

ij<br />

ij<br />

j<br />

∠δ<br />

−δ<br />

i<br />

j<br />

−θ<br />

ij<br />

P<br />

i<br />

=<br />

P<br />

Gi<br />

−<br />

P<br />

Di<br />

=<br />

n<br />

∑<br />

j=<br />

1<br />

VV Y<br />

i<br />

j<br />

ij<br />

cos( δ − δ<br />

i<br />

j<br />

−θ<br />

ij<br />

)<br />

i<br />

= 1,2,..... n<br />

Q<br />

i<br />

=<br />

Q<br />

Gi<br />

− Q<br />

Di<br />

=<br />

n<br />

∑<br />

j=<br />

1<br />

VV Y<br />

i<br />

j<br />

ij<br />

sin( δ<br />

i<br />

− δ<br />

j<br />

−θ<br />

ij<br />

)<br />

i<br />

= 1,2,..... n


Classification of Buses<br />

PQ bus PV bus Slack<br />

bus<br />

- Load bus<br />

- Buses to which only<br />

loads are connected<br />

- Real and reactive<br />

powers are specified<br />

- Load flow solution<br />

determines voltage<br />

magnitude and angle<br />

- Generator bus or voltage<br />

controlled bus<br />

- Buses to which generators<br />

or reactive power sources<br />

are connected<br />

- Real power and voltage<br />

magnitude are specified<br />

- Load flow solution<br />

determines reactive power<br />

and angle<br />

- Reference bus or<br />

swing bus<br />

- A generator bus<br />

having largest output<br />

- Real and reactive<br />

powers are specified<br />

- Load flow solution<br />

determines voltage<br />

magnitude and angle


Numerical Solution Techniques<br />

• Gauss Iterative and Gauss-Siedel methods<br />

• Newton-Raphson method<br />

- Rectangular coordinates<br />

- Polar coordinates<br />

• Fast Decoupled Load Flow method


NRLF method in Polar coordinates<br />

The power flow equations in polar coordinates can be<br />

written as,<br />

P<br />

i<br />

Q<br />

i<br />

=<br />

=<br />

n<br />

∑<br />

j=<br />

1<br />

n<br />

∑<br />

j=<br />

1<br />

VV<br />

i<br />

VV<br />

i<br />

j<br />

( G cosδ<br />

+ B sinδ<br />

)<br />

j<br />

ij<br />

i = 2,..... n<br />

( G sinδ<br />

−B<br />

cosδ<br />

) i = m + 1,..... n<br />

ij<br />

ij<br />

ij<br />

ij<br />

ij<br />

ij<br />

ij<br />

where,<br />

δ<br />

ij<br />

=<br />

δ<br />

i<br />

− δ<br />

j<br />

The correction vector can be written as<br />

∆V<br />

∆δ<br />

n<br />

<br />

∆δ<br />

n<br />

∆V<br />

/ V<br />

<br />

2<br />

m+<br />

1<br />

/ V<br />

n<br />

∂P<br />

∂P<br />

H = L = V<br />

= ∂δ<br />

∂V<br />

∂Q<br />

∂Q<br />

M = N = V<br />

∂δ<br />

∂V<br />

∆P<br />

<br />

∆P<br />

∆Q<br />

n<br />

∆Q<br />

m+<br />

1<br />

<br />

2<br />

n


Computational steps to solve load flow problem<br />

Step 1: Read network admittances, generator, load and<br />

transformer data<br />

Step 2: Form Y bus<br />

matrix<br />

Step 3: Assume initial values of bus voltages except at<br />

slack bus. Initialize iteration count K to zero.<br />

Step 4: Compute P at all buses except slack bus and Q<br />

at all the load buses. Calculate ∆P at all buses<br />

except slack bus, ∆Q at all the load buses.<br />

Step 5: If all the elements of mismatch vectors (∆P, ∆Q)<br />

are within the prespecified tolerance, load flow<br />

solution is achieved. Calculate slack bus power,


Contd.<br />

Step 6: Assemble Jacobian matrix, find its inverse and<br />

calculate correction vector<br />

Step 7: Update angles at all the buses except slack bus<br />

and voltages at all the load buses and go to<br />

step 4<br />

The reactive power limits of generators can be<br />

checked at step 4. If the calculated reactive<br />

power of any PV bus is within the limit, it is<br />

retained as PV bus. If it violates the limit, the<br />

generator output is fixed to the limiting value<br />

and the bus is solved as PQ type for that<br />

particular iteration, with specified value of Q<br />

taken as difference between the limiting value<br />

and the calculated value.


Steady State Model of Wind Farm<br />

• As a negative load (PQ bus)<br />

- assumes a generated real power and a given power factor,<br />

with which the consumed reactive power is calculated.<br />

• As a voltage controlled (PV bus)<br />

- when the voltage at the connection point need to be<br />

controlled at normal voltage by providing reactive power support<br />

• PX bus model<br />

- the real power is known and the reactive power is calculated<br />

as a function of the magnetizing reactance of the generator<br />

• RX bus model<br />

- assumes resistance and reactance of both stator and rotor<br />

as well as the mechanical output power P m<br />

to be known variables


Steady State Model of Wind Farm, Contd.<br />

An improved PQ bus model has been proposed considering the<br />

steady state properties of the induction generator, such that the<br />

reactive power Q is calculated by<br />

Q<br />

≈V<br />

2<br />

⎛<br />

⎜<br />

⎝<br />

X − X<br />

X X<br />

c m<br />

+<br />

c<br />

where, X c<br />

is capacitive reactance<br />

X m<br />

is magnetizing reactance<br />

X is sum of stator and rotor reactances<br />

V is terminal voltage<br />

P is real power of the generator<br />

The improved PQ bus model requires that the specified<br />

reactive power Q be updated after each iteration of the load<br />

flow<br />

m<br />

⎞<br />

⎟ ⎠<br />

X<br />

V<br />

*<br />

A.E. Feijoo and J. Cidras, “Modeling of Wind Farms in the Load Flow <strong>Analysis</strong>”,<br />

IEEE Transactions on <strong>Power</strong> <strong>System</strong>, Vol. 15, No. 1, Feb. 2000, pp. 110-115<br />

2<br />

P<br />

2


PX bus model<br />

• Constant reactance model<br />

- assumes a constant active power generation and a<br />

constant magnetizing reactance, neglecting the stator and<br />

rotor reactances<br />

- constant relation between the voltage level and the<br />

reactive power consumption<br />

• Variable reactance model<br />

- considers saturation of magnetic circuit<br />

- models magnetizing reactance as voltage<br />

dependent


R<br />

( s)<br />

=<br />

Let R(<br />

s)<br />

P<br />

eq<br />

IG<br />

( V , s)<br />

2<br />

r<br />

R<br />

s<br />

+ ( X<br />

= R + R<br />

=<br />

R<br />

s<br />

2<br />

R<br />

r<br />

X<br />

s<br />

m<br />

V<br />

eq<br />

( s)<br />

+<br />

2<br />

m<br />

2<br />

+<br />

( s)<br />

X<br />

X<br />

2<br />

r<br />

)<br />

( s)<br />

2<br />

and<br />

;<br />

X<br />

R(<br />

s)<br />

eq<br />

( s)<br />

=<br />

X ( s)<br />

=<br />

;<br />

Q<br />

X<br />

IG<br />

R<br />

s<br />

r<br />

+<br />

X<br />

s<br />

2<br />

m<br />

2<br />

r<br />

R<br />

s<br />

X<br />

eq<br />

+<br />

( V , s)<br />

=<br />

X<br />

( s)<br />

R<br />

m<br />

+ ( X<br />

2<br />

X<br />

m<br />

r<br />

V<br />

( s)<br />

+<br />

( X<br />

+<br />

2<br />

X<br />

X<br />

m<br />

r<br />

2<br />

)<br />

+<br />

2<br />

( s)<br />

X<br />

r<br />

)<br />

X ( s)<br />

*<br />

N. paensuwan and A. Yokoyama, “Risk-based TTC calculation of a power system<br />

with renewable energy resources”, IEEE <strong>Power</strong> Tech Conference, 2009


Contd.<br />

This PX model introduces a new state variable i.e. the rotor<br />

slip. Therefore, another equation is required to enforce P IG<br />

to its specified value. The power balance equations are<br />

modified as follows.<br />

f<br />

f<br />

Pi<br />

Qi<br />

=<br />

=<br />

( δ , V ) − ( P<br />

( δ , V ) − ( −Q<br />

) = 0<br />

To enforce P IGi<br />

to its specified value, the following equation<br />

is added.<br />

spec<br />

f P IGi<br />

− P<br />

− Q<br />

The augmented power flow problem is<br />

⎡ ∂f<br />

P<br />

⎢<br />

∂δ<br />

⎢<br />

∂f<br />

⎢ Q<br />

⎢ ∂δ<br />

⎢ ∂f<br />

PIG<br />

⎢<br />

⎣ ∂δ<br />

F<br />

F<br />

Pi<br />

Qi<br />

spec<br />

IGi<br />

Di<br />

Di<br />

IGi<br />

( V<br />

( V , s ) = −P<br />

( V , s ) − P = 0<br />

i<br />

∂f<br />

i<br />

∂V<br />

∂f<br />

Q<br />

∂V<br />

∂f<br />

P<br />

P<br />

IG<br />

∂V<br />

∂f<br />

IGi<br />

∂s<br />

∂f<br />

Q<br />

∂s<br />

∂f<br />

P<br />

P<br />

IG<br />

∂s<br />

i<br />

i<br />

IGi<br />

⎤<br />

⎥<br />

⎥⎡∆δ<br />

⎤ ⎡ ∆f<br />

⎥⎢<br />

⎥ ⎢<br />

⎢<br />

∆V<br />

⎥<br />

= ⎢ ∆f<br />

⎥<br />

⎥⎢<br />

⎥ ⎢<br />

⎣ ∆s<br />

⎦ ⎣<br />

∆f<br />

⎥<br />

⎦<br />

Q<br />

P<br />

P<br />

IG<br />

⎤<br />

⎥ ⎥⎥ ⎦<br />

i<br />

, s<br />

i<br />

)) = 0


RX bus model<br />

Equivalent circuit of induction generator<br />

Voltage obtained from each iteration of load flow is used to calculate the slip<br />

2 (1 −s)<br />

P m<br />

= R2<br />

⋅ I<br />

2<br />

⋅<br />

s<br />

where,<br />

I<br />

2<br />

=<br />

−<br />

Vt<br />

[( R1<br />

+ R2<br />

/ s)<br />

+ j(<br />

X1<br />

+ X<br />

2)]<br />

Once, the slip is known, power injections can be calculated<br />

P<br />

g<br />

Q<br />

g<br />

= −V<br />

t<br />

= −V<br />

t<br />

2<br />

2<br />

⋅<br />

( R<br />

1<br />

( R<br />

⋅<br />

1<br />

( R1<br />

+ R2<br />

/ s)<br />

+ R<br />

2<br />

/ s)<br />

+ ( X + X<br />

2<br />

+ R<br />

X<br />

2<br />

m<br />

/ s)<br />

⋅<br />

2<br />

1 2 m 1<br />

2<br />

2<br />

((<br />

R + R / s)<br />

+ ( X + X ) )<br />

1<br />

1<br />

+ ( X<br />

2<br />

+ X<br />

2<br />

)<br />

2<br />

) ⋅(<br />

X<br />

1<br />

+ X<br />

2<br />

+ X<br />

2<br />

)


Solution Techniques of PLF<br />

Numerical Method<br />

•DLF is performed a large<br />

number of times with inputs of<br />

different combinations of nodal<br />

power values<br />

•exact non-linear form of load<br />

flow equations can be used<br />

•Monte Carlo simulation is<br />

used<br />

•time consuming<br />

Analytical Method<br />

•analyzes a system and its<br />

inputs using mathematical<br />

expressions<br />

•uses linearized load flow<br />

equations<br />

•complicated mathematical<br />

computation<br />

•inaccurate due to different<br />

approximations


Probabilistic Model of Wind Farms<br />

• Probabilistic model of the wind speed<br />

Weibull distribution is the best pdf for the description<br />

of a wind speed<br />

1<br />

k ⎛ v − v ⎡<br />

0 ⎞ ⎛ v − v0<br />

f ( v)<br />

= ⎜ ⎟ exp ⎢−<br />

⎜<br />

c ⎝ c ⎠ ⎢⎣<br />

⎝ c<br />

k − k<br />

where, v is wind speed<br />

v 0<br />

is cut-out speed<br />

k is shape parameter<br />

c is scale parameter<br />

⎞<br />

⎟<br />

⎠<br />

⎤<br />

⎥<br />

⎥⎦<br />

*<br />

L. Dong, W. Cheng, H. Bao and Y. Yang, “Probabilistic Load Flow analysis for<br />

power system containing wind farms”, IEEE 2010


Probabilistic Model of Wind Farms, Contd.<br />

• Probabilistic model of wind turbine<br />

P<br />

w<br />

⎧0<br />

⎪<br />

k1v<br />

+<br />

= ⎨<br />

⎪ Pr<br />

⎪<br />

⎩0<br />

P<br />

k<br />

2<br />

v ≤<br />

v<br />

r<br />

vci<br />

≤ v ≤<br />

vr<br />

≤ v ≤<br />

v > v<br />

co<br />

r<br />

k1 = ; k2<br />

= −k1<br />

vr<br />

−vci<br />

v<br />

v<br />

v<br />

ci<br />

r<br />

co<br />

P<br />

P r<br />

0 v ci<br />

v r<br />

v co<br />

Wind Turbine power output<br />

curve curve<br />

v<br />

where, P r<br />

is rated output power<br />

v r<br />

is rated wind speed<br />

v ci<br />

is cut-in speed<br />

v co<br />

is cut-out speed


Probabilistic Model of Wind Farms, Contd.<br />

•Probabilistic distribution of the active power generated by<br />

WT<br />

F(<br />

P<br />

w<br />

w<br />

)<br />

f ( P )<br />

=<br />

=<br />

v<br />

v<br />

ci<br />

∫<br />

co<br />

F<br />

'<br />

f ( v)<br />

dv +<br />

( P<br />

w<br />

)<br />

=<br />

Pw<br />

−k<br />

k<br />

∫<br />

v<br />

1<br />

ci<br />

2<br />

k ⎛ P<br />

⎜<br />

k1c<br />

⎝<br />

f ( v)<br />

dv<br />

w<br />

−<br />

k1v<br />

k c<br />

1<br />

co<br />

⋅ exp<br />

•Probabilistic distribution of the reactive power absorbed<br />

by WT<br />

−<br />

k<br />

2<br />

⎞<br />

⎟<br />

⎠<br />

k −1<br />

⎡ ⎛ P<br />

⎢−<br />

⎜<br />

⎢⎣<br />

⎝<br />

w<br />

−<br />

k1v<br />

k c<br />

1<br />

co<br />

−<br />

k<br />

2<br />

⎞<br />

⎟<br />

⎠<br />

k<br />

⎤<br />

⎥<br />

⎥⎦<br />

f ( Q ) = tan θ<br />

w<br />

w<br />

⋅<br />

k<br />

k c<br />

1<br />

⎛<br />

⎜<br />

⎝<br />

P<br />

w<br />

−<br />

k1v<br />

k c<br />

1<br />

co<br />

−<br />

k<br />

2<br />

⎞<br />

⎟<br />

⎠<br />

k −1<br />

⋅exp<br />

⎡ ⎛<br />

⎢−<br />

⎜<br />

⎢⎣<br />

⎝<br />

P<br />

w<br />

−<br />

k1v<br />

k c<br />

1<br />

co<br />

−<br />

k<br />

2<br />

⎞<br />

⎟<br />

⎠<br />

k<br />

⎤<br />

⎥<br />

⎥⎦


AC Probabilistic Load Flow Model<br />

Linearized Load Flow Model<br />

⎧ w = f ( x)<br />

⎨<br />

⎩ z = g(<br />

x)<br />

;<br />

⎪<br />

⎧ ∆x<br />

= J<br />

⎨<br />

⎪⎩ ∆z<br />

= G<br />

−1<br />

0<br />

⋅ ∆w<br />

= S<br />

Step 1: Input the system data and wind farm data<br />

Step 2: Run the DLF using NR method, so that the<br />

expected values of nodal voltages,line flows,<br />

S 0<br />

, and T 0<br />

are obtained.<br />

Step 3: Compute the cumulants of generation and load<br />

according to their probabilistic distribution<br />

Step 4: Compute the cumulants of the generated active<br />

⋅ ∆ w<br />

power, absorbed reactive power, power<br />

injections and state variables (∆x and ∆z)<br />

Step 5: Obtain the PDF and CDF of ∆x and ∆z<br />

0<br />

⋅ J<br />

−1<br />

0<br />

0<br />

⋅ ∆ w<br />

⋅ ∆ w = T<br />

0


Load Flow with AC-DC <strong>System</strong>s<br />

• Unified method<br />

- solution vector is extended with the dc<br />

variables<br />

- complex to program and hard to combine<br />

with developments in ac power flow solution<br />

techniques such as fast decoupled method<br />

• Sequential method<br />

- ac and dc equations are solved separately<br />

in each iteration<br />

- easy to implement, but convergence<br />

problems are there


Model of Photovoltaic <strong>System</strong>s<br />

PV<br />

P<br />

Array PV<br />

V PV<br />

I PV<br />

Inverter<br />

M, α<br />

P i<br />

, Q i<br />

V i<br />

Three<br />

phase<br />

system<br />

P g<br />

, Q g<br />

Block diagram of grid connected PV system<br />

• Model based on characteristics of PV array<br />

•Model based on characteristics of specific<br />

inverter structure<br />

• Overall PV system model<br />

Y.B. Wang, C.S. Wu, L. Hua and H.H. Xu, “Staedy-state model and power flow<br />

analysis of Grid-Connected Photovoltaic <strong>Power</strong> <strong>System</strong>”,


Model of Photovoltaic <strong>System</strong>s, Contd.<br />

• DC part model<br />

I<br />

cell<br />

=<br />

d(<br />

VI)<br />

dV<br />

I<br />

L<br />

mpp<br />

− I<br />

=<br />

=<br />

0<br />

I<br />

I<br />

= 0<br />

⎡ ⎛ Vcell<br />

+ I<br />

⎢exp⎜<br />

⎣ ⎝ a<br />

mpp<br />

mpp<br />

+ V<br />

+ V<br />

mpp<br />

mpp<br />

dI<br />

dV<br />

aR<br />

cell<br />

sh<br />

mpp<br />

R<br />

s<br />

− I<br />

+ I<br />

⎞ ⎤<br />

⎟ −1⎥ ⎠ ⎦<br />

− V<br />

0<br />

0<br />

cell<br />

+ I<br />

R<br />

sh<br />

cell<br />

R<br />

⎛ Vmpp<br />

+ ImppR<br />

Rsh<br />

exp⎜<br />

⎝ a<br />

⎛ Vmpp<br />

+ I<br />

RsRsh<br />

exp⎜<br />

⎝ a<br />

s<br />

s<br />

mpp<br />

⎞<br />

⎟ − a<br />

⎠<br />

Rs<br />

⎞<br />

⎟ + aR<br />

⎠<br />

s<br />

(1)<br />

(2)<br />

where, I cell<br />

and V cell<br />

are the current and voltage of PV cell<br />

I L<br />

and I 0<br />

are the light current and diode reverse saturation<br />

current<br />

R s<br />

and R sh<br />

are the series and shunt resistance<br />

a is the ideality factor


Model of Photovoltaic <strong>System</strong>s, Contd.<br />

V<br />

I<br />

P<br />

PV<br />

PV<br />

PV<br />

= N<br />

= N<br />

= V<br />

s<br />

pp<br />

PV<br />

N<br />

I<br />

I<br />

ss<br />

cell<br />

PV<br />

V<br />

cell<br />

(3)<br />

(1)<br />

where, I PV<br />

, V PV<br />

, and P PV<br />

are the current, voltage and power of the<br />

PV array<br />

N s<br />

is the series number of PV cell in a PV module<br />

N ss<br />

is the series number of PV modules<br />

N pp<br />

is the parallel number of PV modules<br />

• Inverter part model<br />

2<br />

V<br />

i<br />

= Vi∠α<br />

= MVPV<br />

∠α<br />

; Pi<br />

=<br />

4<br />

P<br />

PV<br />

(4) ; (5)<br />

where, M is amplitude modulation ratio<br />

α is phase shift angle


Model of Photovoltaic <strong>System</strong>s, Contd.<br />

• AC part model<br />

⎥<br />

⎦<br />

⎤<br />

⎢<br />

⎣<br />

⎡<br />

+<br />

+<br />

−<br />

−<br />

=<br />

⎥<br />

⎦<br />

⎤<br />

⎢<br />

⎣<br />

⎡<br />

−<br />

−<br />

−<br />

−<br />

=<br />

⎥<br />

⎦<br />

⎤<br />

⎢<br />

⎣<br />

⎡<br />

−<br />

−<br />

+<br />

+<br />

=<br />

⎥<br />

⎦<br />

⎤<br />

⎢<br />

⎣<br />

⎡<br />

−<br />

−<br />

−<br />

+<br />

=<br />

23<br />

23<br />

12<br />

12<br />

12<br />

12<br />

23<br />

23<br />

12<br />

12<br />

12<br />

12<br />

12<br />

12<br />

13<br />

13<br />

12<br />

12<br />

12<br />

12<br />

13<br />

13<br />

12<br />

12<br />

sin<br />

sin<br />

)<br />

sin(<br />

3<br />

cos<br />

cos<br />

)<br />

cos(<br />

3<br />

)<br />

sin(<br />

sin<br />

sin<br />

3<br />

)<br />

cos(<br />

cos<br />

cos<br />

3<br />

φ<br />

φ<br />

φ<br />

θ<br />

α<br />

φ<br />

φ<br />

φ<br />

θ<br />

α<br />

φ<br />

α<br />

θ<br />

φ<br />

φ<br />

φ<br />

α<br />

θ<br />

φ<br />

φ<br />

z<br />

V<br />

z<br />

V<br />

z<br />

V<br />

V<br />

Q<br />

z<br />

V<br />

z<br />

V<br />

z<br />

V<br />

V<br />

P<br />

z<br />

V<br />

z<br />

V<br />

z<br />

V<br />

V<br />

Q<br />

z<br />

V<br />

z<br />

V<br />

z<br />

V<br />

V<br />

P<br />

g<br />

g<br />

g<br />

i<br />

g<br />

g<br />

g<br />

g<br />

g<br />

i<br />

g<br />

g<br />

g<br />

g<br />

i<br />

i<br />

i<br />

i<br />

g<br />

g<br />

i<br />

i<br />

i<br />

i<br />

(6)<br />

(9)<br />

(8)<br />

(7)


Operation Modes of Grid-connected PV system<br />

There are nine unknown variables of model and seven<br />

independent equations (1), (4-9)<br />

• Mode 1: PV system applies MPPT strategy and unit<br />

power factor. In this mode, (2) is added and Q g<br />

is set to zero. Then V PV<br />

and P PV<br />

can be carried<br />

out by (1) to (3), and other variables can be<br />

solved by (4) to (9)<br />

• Mode 2: PV system applies MPPT strategy and exports a<br />

certain amount of reactive power. In this mode,<br />

the PV system can provide reactive power and<br />

voltage support for power grid.


Input parameters of<br />

power grid, PV system<br />

and meteorology<br />

Set V g<br />

=rated value,Q g<br />

=0, and solve<br />

PV system model to achieve<br />

M,α,V PV<br />

,P PV,<br />

V i<br />

,P i<br />

,Q i<br />

and P g<br />

PQ type<br />

Traditional power<br />

flow analysis<br />

Node type of<br />

PVPCC <br />

PV type<br />

Traditional power<br />

flow analysis<br />

Set the<br />

parameter at the<br />

limit<br />

Yes<br />

Solve PV system model<br />

to achieve M,α,V PV<br />

,P PV,<br />

V i<br />

,<br />

P i<br />

,Q i<br />

,P g<br />

and Q g<br />

If any<br />

parameter<br />

exceeds a<br />

limit No<br />

If the iteration<br />

converges <br />

No<br />

Yes<br />

Output operating<br />

parameters of PV<br />

system and power grid


Thank You

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!