Está en la página 1de 2

Dosano, R. D. et al.

Paper:

A Novel Approach for Determining Distributed


Generations Penetration Level Using Least Square
Minimization (LSM) Curve Fitting
Rodel D. Dosano , Nemia H. Mabaquiao , Godelyn Gallega-Hisole,
Regin A. Cabacas , and In-Ho Ra,
College of Information and Communications Technology, West Visayas State University
Luna Street, La Paz, Iloilo City, Philippines
E-mail: rodel dosano@yahoo.com, rcabacas@wvsu.edu.ph
Department of Electrical Engineering, Iloilo Science and Technology University

Burgos Street, La Paz, Iloilo City, Philippines


E-mail: {drdnemz, godelyn wvcst}@yahoo.com.ph
School of Computer, Information and Communication Engineering, Kunsan National University

558 Daehak-ro, Miryong-dong, Kunsan, South Korea


E-mail: ihra@kunsan.ac.kr
Corresponding author

[Received April 29, 2016; accepted September 1, 2016]

With the publicized benefits offered by renewable 1. Introduction


energy resources, more and more households em-
brace the utilization of stand-alone installations rang- In conventional distribution networks, electric power
ing from small to medium scale systems. In literature, is supplied to the customers from large generating units
several studies provide insights on the effects of inte- commonly fired from fossil fuels, hydrothermal or hydro-
gration of renewable energy (RE) resources to the dis- power sources. Normally the electric power begins from
tribution systems but have inadequacy of considering these generating sources and is distributed to the con-
the penetration levels. Moreover, RE cost reductions, sumers through transmission lines. Consumers in the low
increasing costs of traditional energy sources, and Re- voltage (LV) distribution networks are increasingly in-
newable Portfolio Standards have created the possibil- stalling solar photovoltaic (PV) systems in their homes
ity of significant increase of penetration levels of dis- motivated by increasing benefits from feed-in tariff and
tributed RE generation being installed on distribution issues on environmental concerns [1].
systems. To aid in the evaluation and assist with these In literature, several studies show that different pen-
expansions, new analysis tools are needed. In particu- etration levels and various placements of distributed re-
lar, new RE high-penetration analysis tools and proce- newable resources have different impacts on the distribu-
dures need to be developed and integrated with exist- tion system [24]. The list includes voltage control and
ing conventional methods. This paper presents a sim- stability problems, increased fault duty on circuit break-
ulation based study on distribution system with and ers and protection coordination problems, islanding con-
without integration of RE sources. It takes into ac- ditions, power-quality (PQ) issues, personnel safety, over-
count of the impending effects of these RE integrations voltages, and intermittent [5] or stochastic nature of some
in the distribution system. This paper emphasizes a renewable distributed sources. With these challenges at
novel method of determining the penetration level of hand, this paper presents simulation-based analyses and
Distributed Generation using least square minimiza- models of distribution systems specifically focuses on the
tion (LSM) method. The studies were tested using assessment of the effects of DG penetration levels on the
IEEE 123 bus distribution test feeder and actual data distribution system. Some of the issues that this study
from an existing distribution system to verify the ef- covers include: Effects of unbalanced phase distribution
fectiveness and robustness of the proposed approach. of DGs on the distribution system, which is caused by un-
coordinated massive installations; Voltage violations like
in [6] on some buses and impacts of DGs in systems loss
for different penetration level are also presented.
Keywords: RE resources integration, distributed gener- This study considers balanced and unbalanced config-
ation (DG), DG penetration levels, penetration level mar- urations of the distribution feeders to evaluate parameters
gin like bus voltages, equipment line losses and line flows of
the system. Furthermore, a novel method to determine the
penetration level margin of RE sources from the current

1004 Journal of Advanced Computational Intelligence Vol.20 No.6, 2016


and Intelligent Informatics
Determining DG Penetration Level Using LSM Curve Fitting

Fig. 2. System performance model derived from data plots


using LSM.

tion, it is of most important to include simulations where


the minimum total line loss of the system is explicitly re-
flected.
Modeling of the distribution feeder performance can be
formulated once the system run is completed. This per-
formance model of the distribution feeder can be utilized
to study the impacts of different PV penetration levels as
illustrated in Fig. 2. The data points generated by simu-
Fig. 1. Flow chart of the proposed methodology. lation of PV system integrations are fitted using a least
square minimization (LSM) method to create a perfor-
mance model of the system. Other models considering
different load demands can be determined from the data
operating condition at different load levels is presented. plots from their respective simulations. Furthermore, the
The remainder of this paper is organized as follows: Sec- nature of the data plots which represent the system perfor-
tion 2 discuss the methodology and how the study is im- mance at a given load can best be expressed by a quadratic
plemented and show calculations on how system model- function as:
ing are assumed, Section 3 provides details on the simu-
yn = An x2 Bn x +Cn . . . . . . . . . . (1)
lation environments including actual data for simulation
and the distribution feeders. Section 4 shows the results where yn (x) is the total kW losses of the system; x is
and analyses of the method and lastly, Section 5 presents the variable for the amount of PV system integration; and
concluding remarks. An , Bn and Cn are constants determine using curve fitting
procedures.
The optimum value of x can be found by setting the first
2. Methodology derivative of the quadratic function Eq. (1) to zero.
0 = 2An x Bn . . . . . . . . . . . . . (2)
The main focus of the paper is on the impacts of differ-
ent DGs penetration levels on the distribution feeder sys- The value of x calculated from Eq. (2) determines the min-
tem. The proposed methodology uses an exhaustive ap- imum value of system total power losses (kW) and opti-
proach to translate various possible DGs penetration level mum value of PV penetration level which could be inte-
to get a better insight on its advantages/disadvantages. grated in the distribution system.
This is carried out here through the use of steady state Figure 2 shows the system performance models de-
load flow studies on the distribution feeder. The flowchart rived from simulations with different load demands. Each
of the proposed methodology is shown in Fig. 1. model in the figure shows the total losses in kilowatt
The methodology starts with the initial simulation of (kW) versus installed PV capacity in percentage (%) in
load flow analysis in the distribution feeder using base different load demands respectively. Other system perfor-
case load and generation data. Next, integration of ran- mance models of different system load demand using the
domly selected PV system capacity, bus number and same approach are illustrated in Fig. 3. Individual opti-
phase to the distribution feeder are considered in the sys- mal point of performance of the system is shown in each
tem study. Every complete simulation run of PV integra- curve summaries the ease of using this approach on a sys-
tions on the system updates the feeder data and are saved tem with different load demand. Using this approach, the
as current operating scenario. The number of iterations set available PV capacity to be installed from its current oper-
is distinct to every feeder system to be evaluated. In addi- ating point can be determined. This available PV capacity

Vol.20 No.6, 2016 Journal of Advanced Computational Intelligence 1005


and Intelligent Informatics

También podría gustarte