dc.description.abstract |
In the late 1880s, two exceptionally brilliant inventors, Nikola Tesla and Thomas
Edison embroiled in a battle for establishing the preeminence of alternating current (AC)
and direct current (DC) systems respectively. But due to the numerous advantages of AC
over DC during that period, the latter was forced out of the competition and AC emerged
victorious. Since then, AC system has dominated all the relevant field (viz. generation,
transmission and distribution) of the power system. However, with the recent developments
in the area of power electronic devices and renewable energy technologies, interest in DC
system is revoked. In areas such as long distance transmission, energy storage systems etc.,
DC has undoubtedly proved its edge over the AC system. Power generation, for which AC
was considered to be superior to DC, is now inclining towards DC due to deployment of
renewable energy technologies. On the load side of the power system, electronic loads have
significantly increased in recent times. The result is a tremendous increase in DC power
consuming devices. In the current scenario, there are various advantages of DC over AC
system, but the complete reinforcement of existing distribution system into DC seems to be
impossible at this stage. The reason is, the existing distribution systems is mostly AC. So
researchers are planning a hybrid distribution system (AC-DC distribution system) with both
AC and DC grids.
For proper planning, analysis and optimal operation of transmission or distribution
system, a power flow or load flow study has to be carried out. Therefore, power-flow/loadflow
algorithm for AC-DC distribution system is highly required. However, load flow in
such systems is a challenging task due to non linear characteristics of power converters.
There are well established algorithms for load flow solution of AC transmission systems,
AC-DC transmission systems and for AC distribution systems but for AC-DC distribution
systems till date not much work has been reported. The proposed work aims for the
development of a load flow algorithm for AC-DC distribution network
To acquire the theme objective of this thesis, firstly a power flow algorithm for AC
distribution system has been developed utilising the concept of graph theory and matrix
algebra. The developed load flow methodology is capable of handling any kind of AC
ii
distribution network viz. radial, meshed, single phase and three phase distribution network. It
requires the conventional bus-branch oriented data as the only input. Seven developed
matrices, loads beyond branch matrix (LB), load current matrix (LC), feeder current matrix
(FC), path impedance matrix (PI), path drop matrix (PD), slack bus to other buses drop
matrix (SBOBD), load flow matrix (LFM) and simple matrix operations are utilized to obtain
load flow solutions. In contrast with traditional load flow methods for HVDC systems, the
proposed technique does not require any lower upper (LU) decomposition, matrix inversion
and forward-backward substitution of Jacobian matrix. Because of the aforementioned
reasons, the developed technique is computationally efficient. In case of meshed &/or ring
network they should be converted to equivalent radial network by breaking the loops. The
loop breakpoint matrix/sensitivity matrix has been derived to calculate the net breakpoint
injection to mimic the effect of meshed network in the radial scenario. Thus equivalent radial
network will have to take into account the loop breakpoint injection. The additional
breakpoint injections are reflected in the LB or LC matrix. Once, the meshed network is
converted to radial network, the relevant matrices formulation and load flow procedure are
carried out in same manner as the radial distribution network. For three phase unbalanced
distribution network, the relevant matrices has been modified as per three phase scenario.
The effectiveness of the proposed solution methodology has been tested on several standard
distribution systems. Test outcome reveal efficiency and authenticity of the proposed
research work.
Subsequently, the developed load flow algorithm has been further extended for
solving the load flow problem of AC distribution network with various model of distributed
generations. The mathematical model of DG considered as PQ and PV buses are incorporated
into the proposed algorithm to imitate the injection of DGs in the distribution systems. The
power injected by the DG need to be reflected in the LB matrix of the distribution network.
For the PV type buses or distributed generators, the reactive power generation is adjusted
between the maximum and minimum limits in order to maintain the constant voltage and
constant real power (injected) at the PV bus. The breakpoint matrix has been utilized to
obtain the additional reactive power injection/withdrawal to maintain the specified voltage at
each PV nodes. In the case of weakly meshed distribution network with PV type distributed
generations, the loop breakpoint injections and PV breakpoint injections have been calculated
iii
simultaneously. The net injections is reflected in the LB or LC matrix of the distribution
network. Note that except for some modifications needed to be done for the LC or LB
matrices, the proposed solution techniques require no modification; therefore, the proposed
method can obtain the load-flow solution for AC distribution system in the presence of
distributed generations efficiently. The remaining supporting matrices will reshape
themselves accordingly. The effectiveness of the proposed algorithm has been tested on
several standard distribution systems. Test outcome reveal viability and accuracy of the
proposed research work.
The AC load flow algorithm developed in this thesis work has been modified in a
manner such that it can easily be extended for load flow calculation of AC-DC radial
distribution systems with distributed generations. For our purpose, the AC-DC distribution
network has been subdivided into a number of sub-distribution systems or sub-regions
depending on number of bus-bus interfacing converters present in the distribution system.
Each sub-region will act as separate distribution system. These regions or sub distribution
systems can be interconnected in different hybrid configuration. The relevant matrices
formulation as discussed above has to be carried out for each sub-region separately. The per
unit equivalent model of three-phase PWM AC/DC converter, PWM DC-DC converter and
three-phase AC-DC LCC converter have been proposed in this section. The mathematical
model of DGs considered as P, PQ, PV and Vdc buses are incorporated into the proposed
algorithm to imitate the injection of DGs in the AC-DC distribution systems. The power
injected by the DG need to be reflected in the 𝐿𝐵𝑔 or 𝐿𝐶𝑔 matrix of the AC-DC distribution
network. The sensitivity or breakpoint matrix has been utilized to obtain the additional
reactive power injection/withdrawal to maintain the specified voltage at each PV nodes.
Similarly, when Vdc nodes are present in DC regions of an AC-DC distribution network, the
correct amount of active power injection by the generation units is calculated to compensate
the variance between obtained and specified voltage. The Vdc sensitivity or breakpoint matrix
has been utilized to obtain the additional real power injection/withdrawal to maintain the
specified voltage at each Vdc nodes. Once the net injection by the DGs is calculated, the load
flow solution of active distribution system will be carried out in the same manner as load
flow solution of distribution system without DGs.
iv
Subsequently, the developed load flow algorithm has been further extended for
solving the load flow problem of AC-DC meshed distribution network with various model of
distributed generations. There can exist three kinds of meshes in an AC-DC distribution
network.
(a) Mesh consisting of only AC buses in an AC sub-region is called as type 1 mesh.
(b) Mesh consisting of only DC buses in a DC sub-region is called as type 2 mesh.
(c) Mesh consisting of both AC and DC buses or only DC buses with different voltage levels
are called as type 3 mesh.
The procedure for calculating injected current/injected power will be different for the
different meshed configurations. Suitable procedure has been developed for calculating the
loop breakpoint injections for these three kind of loops in the AC-DC distribution network.
The additional breakpoint injections are reflected in the 𝐿𝐵𝑔 or 𝐿𝐶𝑔 matrix (as per the
injected quantity calculated). Once, the meshed AC-DC network is converted to its
equivalent radial network, the relevant matrices formulation and load flow procedure are
carried out in same manner as the radial AC-DC distribution network. In the case of weakly
meshed distribution network with PV and Vdc type distributed generations, the loop
breakpoint injections and PV breakpoint injections have been calculated simultaneously. Test
results signifies the efficacy and accuracy of the developed load flow algorithm.
In order to prove the applicability of the load flow algorithm, a power flow based
Distribution use of systems (DUoS) charging methodology has been developed to investigate
the impact of protection system reinforcement costs on the consumers associated with
renewable integrated distribution network. Conventional distribution network is a radial
network with a single power source. Usually overcurrent protection schemes are employed
for such system protection for their simplicity and low-cost. With the introduction of
renewable generations, the existing protection coordination needs to be upgraded. Provision
of directional feature and the requirement of high capacity circuit-breakers at certain points
for the protection scheme demands considerable investment. The renovation cost required for
upgrading the protection scheme will significantly impact the network cost requirement and
consequently distribution use of system (DUoS) charges. This thesis works also aims to
investigate the impact of renewable generations on the DUoS charges considering the cost
associated in revamping the protection scheme. A power flow based MW+MVAr-Miles
v
DUoS charging method, that considers used capacity of the network, is proposed to carry out
the DUoS charging calculations. The proposed charging mechanism appraise/penalise the
users in accordance they are affecting system power factor. Accordingly, the proposed
pricing algorithm may encourage users to act based on the economic signal generated at each
location. The proposed charging algorithm has been tested on various standard systems to
examine the impact of renewable generations on the use of network costs. |
en_US |