Professional Documents
Culture Documents
MAHAVIDYALAYA
• where P and Q denotes real and reactive power. At each bus the generator and
load power can be combined so that the net 3-phase complex power
flowing into ith bus can be written as:
Si = SGi – SLi = (PGi – PLi) + j(QGi – Q Li) = Pi + jQi (2)
(4)
Equation (4) can be written as:
where (5a)
Y11 = y10 + y12 + y13 + y14
Y22 = y20 + y12 + y32
Y33 = y30 + y13 + y23 + y34
Y44 = y40 + y14 + y34
Y12 = Y21 = -y12
Y13 = Y31 = -y13
Y14 = Y41 = -y14
Y23 = Y32 = -y23
Y34 = Y43 = -y34
Eq. (5a) can be written in compact form as:
For an n bus system Ibus is an n×1 vector with general entry Ii , Vbus is an n×1 vector
with general entry Vi
• We add the elements one by one and modify the entries of [Ybus] as per
eq.(7).
• If an element is connected from ith bus to reference, only entry is
Formation of Ybus and Z bus
[Using Singular Transformation]
Fig. 3
Primitive Network
Any network consists many branch having active and/or passive element in
it.
Like Fig. 4(a),(b)
Iip+Jip= yipVip
Jip= yipEip
yip= 1/zip
(1). Line resistance are neglected which means that the active power
loss in the lines is zero. This reduces the complexity of equation
because the total active power generation become equal to the total
active power demand.
(2) The angle 𝛿i is small so that sin𝛿i=𝛿I, that approximation converts
the non linearload flow equation into linear ones so that analytical
solution is possible.
(3). All buses except the swing bus are voltage controlled buses.this
means that the voltages of all the buses are specified.
The only unknows are angle 𝛿 for slack bus 𝛿i=0 THUS, We have n-1
linear equation which the value of 𝛿 at all can be found.
GAUSS-SEIDEL METHOD-SOLUTION OF
ALGEBRAIC EQUATIONS
A simultaneous non linear algebraic equatins cannot be solved the usul
method for the solution of these equation iterative techniques are necessary
• The values of Ki and Lip are computed once in the beginning and then used in every iteration.
• It is necessary to assume initial values of V2, V3, ... Vn. We know that the voltages at different buses in a
power system are close to 1 PU. Therefore, we assume V2, V3, ... Vn to be 1+ j0 initially. At every step of
iteration we use the most updated values of the bus voltages to compute the new values of the bus voltages.
When (k+ 1)th iteration is being performed to calculate Vi, the most updated values of the bus voltages are
the values obtained in (k+1)th iteration for all buses previous to ith bus (i.e., for p = 1,2... i-1) and the values
obtained in kth iteration for ith bus and the buses after the ith bus (i.e., for p= i, i+1 ... n). In view of this
observation Eq. (10.33) can be written as under for the (k+1) iteration
• The iterative process is continued till the changes in the bus voltages in the successive iterations become less
than a certain pre-specified tolerance level for all buses, i.e.,
• To reduce the number of iterations, an acceleration factor ‘a’ can be used. a=1.6 is considered to be a good
value for load flow studies.
• When all bus voltages have been determined S, is obtained from Eq. (10.22b) and then conjugate of S', gives
Si.
• The final step in load flow study is the calculation of line flows. A line connecting ith and pth buses. The line
is represented by a ‘pi' circuit. The current I and complex power S can be found as
• Similarly, the power fed by pth bus into the line connecting pth and ith buses is
• Thus, power flow over all the lines can be computed.
• The power loss in the line connecting ith and pth buses is given by S +S and the total transmission loss is the
sum of losses over all the lines.
Modification of GS method when PV buses are present:
• Some of the buses in a power system are voltage controlled buses where P and V are specified and Q and
'delta' are unknowns. It is necessary to modify the above computational procedure. Let the buses be
numbered as
• i= 1 slack bus
• i= 2, 3, ... m PV or voltage controlled buses
• i= m+1, m+2, ...n PQ buses
• For the voltage controlled buses, the bus voltage must be equal to the specified voltage |V|. Moreover the
maximum and minimum reactive powers at these buses are also specified and the value of Q, (for i=2,3 ... m)
must lie between these limits. Thus, the conditions to be met with are
• The second requirement may be violated if the specified bus voltage |V| is either too high or too low. It is to
be noted that it is possible to control |V| only by controlling Q. If during any iteration we find that Q is
outside the limits specified, Q is fixed at the minimum or the maximum value as the case may be and then
that bus is treated as a PQ bus. In view of these observations the procedure of calculation for PV buses is as
under:
(a) Calculate Q using Eq. (10.24b) which is reproduced below
• At the start of (k + 1)th iteration, |V| , i.e., the magnitude of V obtained during kth iteration may not
necessarily be equal to |V|. We save this value of V.
• The values of |V| and 'delta' to be used in Eq. (10.39) are those obtained during (k + 1)th iteration (for p =1
to i-1) and those obtained during kth iteration (for p=i to n). Moreover for every iteration |V| must be set
equal to |V|. In view of this Eq. (10.38) can be written as:
(b) Check Qi to see if it is within limits, i.e., Qi,min <Qi <Qi,max. If so calculate new value of K, and then use Eq.
(10.34) to compute V using |V| and 'delta' for the magnitude and phase angle of V. Reset |V| to |V|, but retain
the phase angle ‘delta’ and continue to the next bus.
• If Qi> Qi,max, set Qi = Qi,max, find new value of K and treat the ith bus as PQ bus and continue the computations
as in section (10.9.1). Similarly, if Qi< Qi,min, set Qi = Qi,min, find new value of K and treat the ith bus as PQ bus
and continue the computations as in section (10.9.1).
• Newton Raphson
Method
• Newton Raphson method is used to solve a system of non linear algebraic equations of the
form f(x)=0
• Now consider a set of n non linear algebraic equation given by
fi(x1,x2,…….,xn)= 0
(1)
Where i= 1,2,3,……,n
• Let X10,x20,……..xn0
be the initial guess of unknown variab
• Δx10, Δx20,…….. Δxn0
be the respective correction therefore,
f (x 0+ Δx 0,x 0+ Δx 0,…….,x 0+ Δx 0)= 0 i=0,1,2,…..n (2)
• Now expand above equation using Taylor’s series
• If the higher order terms are neglected then equation 3 can be written in matrix form
•
(4)
• In vector form equation 4 can be written as
F0+J0ΔX0 = 0
Or F0 = - J0ΔX0
(a) Sparsity: In a large power systems each bus is connected to only a small number of
other buses (may be two or three). So, bus admittance matrix of a large power system is
very sparse (i.e., it has a large number of zero elements). The sparsity feature of Y bus
matrix extends to Jacobian matrix also. The sparsity of an n×n matrix is defined as
Sparsity = × 100
In a large power system sparsity may be as high as 97%. It may be noted that through Y bus
is sparse, Zbus is full.
Modern load flow studies take advantages of sparsity to reduce computer memory
requirement and save computer time.
(b) Compact storage scheme
In a compact storage scheme only non-zero elements are stored and it is ensured that full
position identification can take place. Consider a sparse matrix
YBUS=
The vector JCOL identifies the column position of each elements in the vector YOFFD,
e.g., the elements -12 is in second column, -8 is in the fourth column. -12 is in the first
column, -10 is in fourth column and so on.
(c) Gauss elimination, Triangular factorization
The load flow solution requires obtaining inverse of a matrix. When a matrix is sparse,
the matrix inversion is a very inefficient approach.
Instead, it is better to use Gauss elimination which is uses triangular factorization and
back substitution. In triangularization the off diagonal elements in each row below the
main diagonal are made zero and the diagonal elements of each row is normalized as
soon as the processing of that row is completed.
After triangularization it is easy to obtain the solution by back substitution.
Step 2 : Divide the elements of third row by a21 (i.e., -8.868) and subtract it from first row.
The result is
Step 3 : Divide the elements of third row by a31 (i.e., 2.096) and subtract it from first row.
The result is
Continue…
Step 4 : Divide the elements of second row by the new value of a22 (i.e., 1.5074). The result
is
Step 5 : Divide the elements of third row by new value of a32 (i.e., 1.5074) and substact it
from second row. The result is