Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

GUCON

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

2021 IEEE 4th International Conference on Computing, Power and Communication Technologies (GUCON)

University of Malaya, Kuala Lumpur, Malaysia. Sep 24-26, 2021

Steady State Analysis of Distribution System With


Wind and Solar Under Various Electric Vehicle
2021 IEEE 4th International Conference on Computing, Power and Communication Technologies (GUCON) | 978-1-7281-9951-1/21/$31.00 ©2021 IEEE | DOI: 10.1109/GUCON50781.2021.9573903

Load Models
Nasir Rehman M. D. Mufti Neeraj Gupta, SMIEEE
Department of Electrical Engineering Department of Electrical Engineering Department of Electrical Engineering
National Institute of Technology National Institute of Technology National Institute of Technology
Srinagar, India Srinagar, India Srinagar, India
nasirrehman92@gmail.com muftimd@yahoo.com neeraj5822@gmail.com

Abstract—This work proposes a load flow analysis based effect on grid stability and power quality. The increased
on backward and forward sweeps to evaluate the effects of penetration of electric vehicles in the network due to growing
electric vehicle (EV) loading on a distribution system powered energy demand has resulted in the additional challenges in the
by distribution generation technologies like solar and wind.
The simulation has been done on an IEEE-33 bus system. distribution network. EV charging load requirement and EV
The simulation outcomes demonstrates the impact of integrating load characteristics vary from those of other conventional net-
distribution generations at the grid and EV load models with work loads. The site, schedule, and duration of EV charging,
impact levels ranging from low to high for voltage dependent along with its active and reactive power demand, cannot all be
loads (EV2), constant current load (EV3) and polynomial load forecast ahead and thus may compromise in grid constraints
models (EV1), respectively.The study explores the various issues
of incorporating EV loads with renewable energy sources (RES) locally and regionally.
and evaluates proposed solutions for their proper integration in Various effective power flow approaches have been devel-
distribution network. The study also assesses several modelling oped in the research [1] - [3] to ascertain the vulnerabilities
sections of the studied system in order to assist researchers in throughout the distribution systems. The optimal positioning
modelling these kind of systems for research purposes, as well as well as sizing of DGs have been thoroughly presented
as to comprehend and quantify the constraints for integration
techniques. in the literature to ascertain how all these aforementioned
Index Terms—Electric vehicle, photo-voltaic system, wind- disturbances directly and indirectly affects the power flow
turbine generating system, backward-forward sweep load flow and network security of the system. Different optimization
method techniques have been used to enhance the voltage profile and
system reliability of integrated distribution networks. Genetic
I. I NTRODUCTION algorithm [4], ant colony optimization [5], particle swarm
Renewable energy sources have gained a lot of interest optimization [6], cuckoo search [7], and other various algo-
over the years and have became indispensable part of the rithms to determine the best nodes/buses for optimal sitting
integrated grid. They are environment friendly as well as and sizing of DGs. These load flow methodologies are based
offer consumers a cost-effective alternative in view of the either on mathematical or empirical analysis [4], [8] - [11].
escalating costs of conventional fossil fuels. The erratic role Monte-carlo simulation is widely used [11] in mathematical
of renewable energy type distributed generations (DGs) like assessments of the effects of DGs on distribution systems.
wind and solar as well as the incorporation of electric vehicle However, it is time intensive due to the assessment of every
(EV) storage into the distribution network is continuously simulation. In empirically derived point estimation methods
growing with each day, triggering significant complexities and (PEM), initially the weights and points are derived by PEM [1]
variabilities. An extensive understanding of the impacts of followed by the proportional percentage of cumulants of input
DGs and load variability in distribution network is crucial for variables and then finally the cumulants and output instants are
efficient renewable energy accommodation. calculated. With the incorporation of wind turbine generating
A marginal increase in DGs in an inadequate distribution systems (WTGS), multiple PEM-based techniques have been
network affects voltage profile and has also a significant effect explored in [3] and [8]. An asymmetrical two point estimate
on the security and reliability of the distribution network. The approach is presented in [12] to address the complexities
wind and PV conversion systems are uncertain in their output in distribution system with photo voltaic (PV) and wind
powers because of the inevitable and deterministic nature of vulnerabilities. In [13], WTGS for real power injection is
solar irradiance and wind velocities. The inconsistent and illustrated as an ambivalent with a beta distribution for wind
dispersed presence of wind and solar energies into grid adds speed, while rayleigh distribution and weibull distribution
new references of ambiguities that may have a detrimental are used in [8], [14]. The comprehensive wind generator

978-1-7281-9951-1/21/$31.00 ©2021 IEEE 1

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY SRINAGAR. Downloaded on January 09,2023 at 10:43:54 UTC from IEEE Xplore. Restrictions apply.
studies for determining reactive powers are mentioned in [13]. impedance. Then the voltage at the ith node at kth iteration is
Probabilistic confined power flow calculations are used to computed as:
model the charging loads [15] for EVs in which queuing theory
(Vi )k = (Vi−1 )k − Zi (Ii )k (1)
[16] is used to determine the overall charging-discharging
power of EVs. As EV charging consumption is probabilistic, B. Backward sweep
and the number of EVs incorporated in the distribution system The backward sweep computations depends on KCL that is
increasing steadily due to environmental issues, the power used to compute the power flow and the branch currents with
system’s complexity grows. the viable node voltages to form BIBC matrix. The power at
Thus a satisfactory power flow method is needed that can any ith bus is the total load connected on the system beyond
address any form of complexities efficiently and precisely take ith node and is given as:
into account significant inclusions of WTGS, PVs and EV load
models into the grid. To properly address the aforementioned Si = Pi + jQi (2)
problems in an integrated distribution network, a load flow
Pi as active power and Qi as reactive power are at ith node.
method acquired from bus injection to branch currents (BIBC)
The branch current computed through backward sweep is
and branch current to bus voltage (BCBV) technique has been
given by:
formulated based on a specific class of backward-forward  ∗
sweep (BFS) method, that will achieve consistent results by Si
(Ii )k = − Yi (Vi )k−1 (3)
restricting the number of simulations needed and improves (Vi )k−1
the voltage profile of the system while decreasing its overall When DGs and EV loads are incorporated with the grid
losses. To simplify, the following are among the paper’s most at preferable nodes to enhance the voltage profile and grid
significant contributions: stability, the BIBC and BCBV power flow method prompts
• Three different forms of electric vehicle loads are mod- certain modifications to the algorithm which are illustrated
elled to address the challenges with a steady state analysis case wise as;
of integrated distributed networks. 1) When the WTGS is connected to dispatch power into
• The challenges in radial distribution system using power
the grid: The active power is introduced as negative (-P)
flow analysis could be used to model the electrical at node which implies it is generating and the reactive
networks and carry out appropriate planning of RES for power as positive (+Q) signifying that WTGS needs
EV development in future. power.
The entirety of this paper is organised as follows: In section
II, The power flow analysis implemented is discussed for an Pi,wnd = Pi − Pwnd
integrated distribution network. Mathematical modelling of Qi,wnd = Qi + Qwnd
WTGS and PV are explained along with three different forms
2) In PV, the active power obtained is explicitly inserted at
of EV loads in section III. The overview of methodology
its buses and the reactive power remains zero.
in section IV and its impact on an integrated distribution
networks with EV loads is discussed in section V. Section Pi,P V = Pi − PP V
VI covers the outline and future context of this research work. Qi,P V = Qi
3) EV load model establishes the VDL characteristics. The
II. BACKWARD – FORWARD SWEEP LOAD FLOW METHOD
modifications in both active and reactive power flows to
The BFS algorithm is an iterative approach wherein two the system are given as:
sequential steps are carried out for load flow analysis. In the
first step, power flow is calculated through the branches in Pi,EV = Pi + PEV
the backward manner starting from the last node towards the Qi,EV = Qi + QEV
source node and in the next step of recursive iterations, the
voltages at each nodes are calculated from the source node III. M ODELLING
upto the last node in a forward manner.
A. Modelling of WTGS
A. Forward sweep WTGS harnesses the energy contained in the wind and con-
The forward sweep is the calculation of voltage drops verts it into electrical energy. Wind generators are described
by using branch currents. This approach is based on KVL, by how they operate when configured with the grid. The
where the node voltages are obtained in the forward manner wind turbine employs the steady state equivalent model of
proceeding with the branch currents to form BCBV matrix. an induction machine [17]. Fig. 1 illustrates the equivalent
At each node, the voltages are computed by the variance circuit of an induction machine. The wind system discussed
between the preceding node voltage and voltage drop in the in this section is stall-regulated fixed speed wind turbine to
subsequent branches. Let the voltage at any ith node as (Vi ) determine the power output for a given terminal voltage and
and the corresponding branch current as Ii with Zi as branch wind speed. The active power Pg and reactive power Qg of

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY SRINAGAR. Downloaded on January 09,2023 at 10:43:54 UTC from IEEE Xplore. Restrictions apply.
induction machine are illustrated by equations (4) and (5),
respectively [17]. The WTGS is directly interfaced to the
power system network to feed generated powers on buses
through transformers.

Fig. 2. Representation of solar cell circuit

output voltage and output current of PV cells using equation


Fig. 1. Steady state equivalent model of induction machine (6).

(K + Hs + Ls2 ) C. Modelling of EV loads


Pg (V, s) = −V 2 ∗ (4)
(C − Ds)2 + (E + F s)2 EVs are modelled in numerous forms in radial distribution
(A + Bs2 ) system. In this paper, load flow analysis incorporating EV
Qg (V, s) = −V 2 ∗ (5)
(C − Ds)2 + (E + F s)2 loads are addressed through VDL modelling. Three different
forms of EV loads are taken into consideration to explore
where, V denotes the terminal voltage and s is the machine’s
load flow analysis by overcoming the issues with steady state
slip. A to L are constants given as:
analysis. Polynomial load is a first form of EV load model
(EV 1) in which active and reactive powers are obtained as
A = R22 (X1 + Xm ) [19];
B = (X2 + Xm )[X2 Xm + X1 (X2 + Xm )]      
2
Vi Vi
C = R1 R2 , D = X2 Xm + X1 (X2 + Xm ) PEV 1 = Pr p3 + p2 + p1
Vr Vr
E = R2 (X1 + Xm ), F = R1 (X2 + Xm )       (7)
2
2 Vi Vi
H = R2 Xm , K = R1 R22 , L = R1 (X2 + Xm )2 QEV 1 = Qr q3 + q2 + q1
Vr Vr
where, Xm is magnetizing reactance, R1 as stator resistance
and R2 as rotor resistance. The second form is a VDL based for EV load model (EV 2).
In Appendix A, the FSWG parameters are stated. In this type, the equations obtained from the battery and fast
charging station modelling confines the active and reactive
B. Modelling of Photo Voltaic System
powers [20]. The active power obtained and the power factor
The modelling of PV generation requires only steady state of the charging station represents the reactive power of model.
output performance of system and doesn’t need to undergo   2 
the complex characteristics of the control system structural Vi
adjustment programme. This section outlines a mathematical PEV 2 = Pr b + a
Vr (8)
analysis for determining the output power of PV modules.
Solar PV systems incorporate PV cell arrays, control modules, QEV 2 = PEV 2 ∗ tan θ
inverters and certain other sub-assemblies. The output power Constant current load model is the third category of EV load
of PV can be computed by using solar radiations accessible (EV 3) which is primarily based on the exponential load model
on the inclined surface, the ambient temperature, and the data of general load. The reactive power is determined as zero
provided by the panel manufacturer. The PV system modelling (unity power factor) [21] and the active power is computed
is summarized by the circuit diagram depicted in Fig. 2 The as:
output current generated by solar irradiance is defined as solar  alphaaEV
PV current Ipv, which is governed by the equation as follows: Vi
PEV 3 = Pr
Vr (9)
     
V + Ipv Rs V + Ipv QEV 3 = 0
Ipv = Ig − Io exp −1 − (6)
vt n Rs The models discussed above are used to evaluate the grid
where, Ig is the photon current proportionate to solar insola- effects under balanced load. The Modelling of EV loads are
tion and Io is the reverse saturation current of diode obtained developed at nominal load voltages.
by diode’s shockley equation under specific temperature. The parameters of all three EV load models are stated in
The solar photovoltaic output power is calculated using the Appendix B.
values of the parameters specified in [18] by computing the

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY SRINAGAR. Downloaded on January 09,2023 at 10:43:54 UTC from IEEE Xplore. Restrictions apply.
IV. M ETHODOLOGY
There are several methodologies of load flow studies in Start

distribution system. The vulnerabilities and instability of sys-


tem increases when DGs are incorporated. In this work, we Review the network
and identify the
have used power flow method based on backward and forward nodes for numbering
sweep method on an IEEE-33 bus system integrated with DGs
and EV charging load models that addresses nearly all the
complexities of integrated grid efficiently and precisely. Initialize bus
and line data
The method is based on branch currents and node voltages.
The power flow of the system is computed by calculating
the branch currents in a backward manner which is formed Update Bus
data
as BIBC matrix and the node voltages in a forward manner
to form BCBV matrix. The algorithm recursively iterates the
BIBC and BCBV matrices to monitor the power flow of the Form BI BC/
Add EV loads
at 4 different
system and procure the network losses. BCBV Matrix
nodes
steps of algorithm of BIBC and BCBV load flow method
• Step 1: Review the network so that numbering of Obtain node voltage , Integrate
branches is done. branch currents, and
compute power flow
wind/PV with
them
• Step 2: Read the line data and bus data of the network.
• Step 3: Form BIBC and BCBV matrices by using
equation’s (3) and (1) respectively, and compute branch NO
Identify weak
0.91<Vset<1.05
currents and node voltages. nodes

• Step 4: Check node voltage limits within the set reference Yes
voltage (in this case,0.925 < Vset < 1.05) and identify
Print node
weak nodes. If voltages at nodes are not violated, then voltages and
power loss
go to step 6.
• Step 5: Integrate DGs (WTGS and PV) at weak nodes
Stop
alternatively along with EV charging load models at
random nodes and accordingly update bus data in step
2 and repeat the steps. Fig. 4 and Fig. 5 shows the DGs Fig. 3. Flowchart of algorithm
in the system
• Step 6: Calculate line flows and power flows using branch
currents and node voltages calculated in step 3, and
monitor the voltage profile of the network.
The proposed algorithm is shown in Fig. 3 in form of
flowchart.

V. R ESULTS AND D ISCUSSION


In this study, an IEEE 33 test bus radial distribution system
is taken for analysis of load flow. The test system is having
33 nodes and 3 end-nodes, operated at 12.66 KV. The total
load connected at 33 nodes is 3.72MW and 2.3 MVAR. The
voltages at all nodes are computed by forward sweep algorithm
at reference voltage of 1.0 p.u. The total active and reactive
losses without DGs and EV loads incorporated in the system
are computed by backward sweep are shown in Fig. 6 and Fig.
7.
Two WTGS are incorporated at two preferable nodes to
analyze the enhancement in voltage profile and reduction in
losses and then alternatively, four PV systems are incorporated
at two weak nodes and at two end nodes to verify the effects.
The foremost impacts of WTGS and PV systems upon
grid is encapsulated in Fig. 8 and Fig. 9, respectively. The
Fig. 4. WTGS and EVs incorporated in 33 bus distribution system
voltages at weak buses are enhanced and the overall system

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY SRINAGAR. Downloaded on January 09,2023 at 10:43:54 UTC from IEEE Xplore. Restrictions apply.
Fig. 7. Losses of system when PV and EV loads are incorporated

Fig. 5. PV and EVs incorporated in 33 bus distribution system

Fig. 8. Voltage profile of system when WTGS and EV loads are incorporated
losses are reduced and are clearly depicted in Fig. 6 and Fig. 7.

Fig. 6. Losses of system when WTGS and EV loads are incorporated


Fig. 9. Voltage profile of system when PV and EV loads are incorporated

Due to unpredictable and voltage dependent EV loads,


the test bus system is incorporated with all the three above
discussed EV load models at 4 different nodes and the impacts
of such voltage dependent loads on the distribution system are
encapsulated in Fig. 8 with WTGS and in Fig. 9 with PV. The
representation of EV loads with WTGS and PV in the system
are shown in Fig. 4 and Fig. 5 respectively
Fig. 10 illustrates when both WTGS and PVs are incorpo-
rated along with EV loads to analyze the impact on distribution
system. Two WTGS and four PV system are incorporated
at preferable and end nodes with all the three EV load
models connected at different nodes. The overall system loss
is reduced and the voltage profile of the system is very much
enhanced and maintained within the acceptable limits. Fig. 10. Voltage profile of system when WTGS, PV and EV loads are
incorporated

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY SRINAGAR. Downloaded on January 09,2023 at 10:43:54 UTC from IEEE Xplore. Restrictions apply.
From this analysis, it has been observed that with the incor- [3] C. Chen, W. Wu, B. Zhang, and H. Sun, “Correlated probabilistic load
poration of EV, the voltage profile of the system dips, which flow using a point estimate method with nataf transformation”, Int. J.
of Electr. Power and Energy Syst., vol. 65, pp. 325–333, Feb. 2015.
can be improved by incorporating the RES in the distribution [4] I. Tegani et al., “Optimal sizing study of hybrid wind/PV/diesel power
system. generation unit using genetic algorithm,” 4th International Conference
on Power Engineering, Energy and Electrical Drives, 2013, pp. 134-140.
VI. C ONCLUSION [5] L. Wang and C. Singh, “Reliability-Constrained Optimum Placement of
Reclosers and Distributed Generators in Distribution Networks Using an
In this study, the various EV load models along with RES Ant Colony System Algorithm,” IEEE Transactions on Systems, Man,
has been integrated in the distribution system. The algorithm is and Cybernetics, Part C (Applications and Reviews), vol. 38, no. 6, pp.
757-764, 2008.
carried out in the radial distribution network that is affected by [6] H. Musa and S. S. Adamu, “Enhanced PSO based multi-objective
the incorporation of wind, PV and EV uncertainities leading to distributed generation placement and sizing for power loss reduction and
irregular voltage profile and overall network losses. The test voltage stability index improvement,” 2013 IEEE Energytech, 2013, pp.
1-6.
has been conducted on an IEEE-33 bus system. The results [7] M. M. fard, R. Noroozian, and S. Molaei, “Determining the optimal
demonstrate that integration of DGs and EV load models placement and capacity of DG in intelligent distribution networks under
discussed have significant impacts on the distribution network. uncertainty demands by COA,” Iranian Conference on Smart Grids,
2012, pp. 1-8.
From the analysis, following conclusion can be drawn: [8] X. Ai, J. Wen, T. Wu, and W. J. Lee, “A discrete point estimate method
• With the incorporation of EVs, there is dip in the voltage for probabilistic load flow based on the measured data of wind power”,
IEEE Trans. on Ind. Appl., vol. 49, no. 5, pp. 2244–2252, Sept. 2013.
profile of the system. [9] J. Usaola, “Probabilistic load flow with wind production uncertainty
• The voltage profile of the system can be improved and using cumulants and cornish- fisher expansion”, International Journal of
the losses can be minimised by integration of RES in the Electrical Power and Energy Systems, vol. 31, no. 9, pp. 474–481, Oct.
2009.
distribution system. [10] B. R. Prusty and D. Jena, “A critical review on probabilistic load
• With the proper coordination of EV and RES, we can flow studies in uncertainty constrained power systems with photovoltaic
have a robust system. generation and a new approach”, Renewable and Sustainable Energy
Reviews, vol. 69, pp. 1286–1302, March 2017.
The future scope of this work concerns the challenging [11] Y. Wang, N. Zhang, Q. Chen, J. Yang, C. Kang, and J. Huang,
issue of energy management in case of high electric vehicle “Dependent discrete convolution based probabilistic load flow for the
active distribution system”, IEEE Trans. on Sust. Energy, vol. 8, no. 3,
penetration. pp. 1000–1009, July 2017.
[12] A. Soroudi, M. Aien and M. Ehsan, “A probabilistic modeling of
A PPENDIX photo voltaic modules and wind power generation impact on distribution
networks”, IEEE Syst. J., vol. 6, no. 2, pp. 254 –259, June 2012
A. WTGS parameters: [13] Neeraj Gupta, “Probabilistic load flow with detailed wind generator
models considering correlated wind generation and correlated loads”,
Renewable Energy, vol. 94, pp. 96–105, June 2016.
R1 = 0.0027Ω, R2 = 0.0022Ω [14] D. Villanueva, J.L. Pazos, and A. Feijoo, “Probabilistic load flow
including wind power generation”, IEEE Trans. on Power Syst., vol.
X1 = 0.025Ω, X2 = 0.046Ω 26, no. 3, pp. 1659–1667, Aug. 2011.
[15] J. G. Vlachogiannis, “Probabilistic constrained load flow considering
Zm = j1.38Ω integration of wind power generation and electric vehicles”, IEEE Trans.
Rated voltage Vr = 690V on Power Syst., vol. 24, no. 4, pp. 1808–1817, Nov. 2009.
[16] U. N. Bhat, An Introduction to Queueing Theory: Modeling and Anal-
Rated Power Pr = 900KW ysis in Applications, Birkhäuser Boston Springer Science, New York,
2008..
B. EV load parameters: [17] T.Petru and T.Thirnger, Modelling of wind turbines for power system
studies, IEEE Trans. Energy Convers. Vol. 17, N¡4, pp. 1132-1139, Nov.
p1 , p2 and p3 = Active power coefficients equal to 2002.
−0.1773, 0.9949 and 0.1824 of power constant, current [18] H.Bellia, R.Youcef, M. Fatima, A detailed modeling of photovoltaic
module using MATLAB, NRIAG Journal of Astronomy and Geophysics.
constant, and impedance constant respectively. Volume 3, Issue 1, June 2014, Pages 53-61.
q1 , q2 and q3 = Reactive power coefficients equal to 4.993, [19] A.M.A. Haidar, K.M. Muttaqi, Behavioral characterization of electric
−12.910 and 8.917 of power constant, current constant, and vehicle charging loads in a distribution power grid through modeling of
battery chargers, IEEE Trans. Ind. Appl. 52 (1) (2016) 483–492.
impedance constant respectively. [20] C.H. Dharmakeerthi, N. Mithulananthan, T.K. Saha, Impact of electric
alphaaEV = Exponential index with a power constant of 1. vehicle fast charging on power system voltage stability, Int. J. Electr.
θ = Angle between active and reactive power. Power Energy Syst. 57 (2014) 241–249 2014/05/01.
[21] F. Milano, O. Hersent, Optimal load management with inclusion of
a = exponential constant equal to 0.007. electric vehicles and distributed energy resources, IEEE Trans. Smart
b = Power constant equal to 0.93. Grid 5 (2) (2014) 662–672.

R EFERENCES
[1] J. Usaola, “Probabilistic load flow in systems with wind generation”,
IET Gener. Transm. and Distrib., vol. 3, no. 12, pp. 1031 –1041, Dec.
2009.
[2] Pei Zhang and S.T. Lee, “Probabilistic load flow computation using
the method of combined cumulants and gram-charlier expansion”, IEEE
Trans. on Power Syst., vol. 19, no. 1, pp. 676 – 682, Feb. 2004.

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY SRINAGAR. Downloaded on January 09,2023 at 10:43:54 UTC from IEEE Xplore. Restrictions apply.

You might also like