OpenDSS-Based Distribution Network Analyzer
in Open Source GIS Environment
R. Gonza´lez, Student Member, IEEE, A. Arguello, Graduate Student Member, IEEE,
G. Valverde, Member, IEEE, and J. Quiro´s-Torto´s, Member, IEEE
Electrical Power and Energy Research Laboratory
School of Electrical Engineering, University of Costa Rica
11501-2060 UCR, San Jose´, Costa Rica
gvalverde@eie.ucr.ac.cr, jquiros@eie.ucr.ac.cr
Abstract—The adoption of low carbon technologies (e.g., pho-
tovoltaic systems) is expected to increase in the near future
given their contribution to reduce greenhouse gas emissions.
However, high penetrations of these new technologies are likely
to result in technical problems on the distribution networks.
To truly understand these impacts, nonetheless, utilities need to
have detailed models of their networks and advanced simulation
tools, so studies can then be performed. This paper presents
an OpenDSS-based distribution network analyzer in an open
source GIS environment that enables engineers to easily carry out
complex network analyzes ranging from snapshot to harmonic
power flows. Critical to facilitate its use, the simulation tool needs
limited input information to undertake these detailed studies.
Illustrations on a real distribution network with more than 10,000
customers demonstrate the efficiency of the analyzer to perform
studies, display and store the corresponding results.
Index Terms—Distribution network analyzer, distribution net-
work, low carbon technologies, OpenDSS, Python, QGIS.
I. INTRODUCTION
The increasing adoption of photovoltaic systems, electrical
energy storage systems and electric vehicles will contribute in
reducing greenhouse gas emissions - the main cause of climate
change [1], [2]. Although the uptake of these technologies is
expected to create new opportunities for the electricity system,
high penetrations are likely to result in significant technical
problems on the distribution networks (e.g., thermal overloads,
high/low voltages, voltage unbalance, harmonics) [3], [4].
To truly quantify the corresponding benefits and/or impacts
on the Medium Voltage (MV) and Low Voltage (LV) distri-
bution circuits, it is, therefore, critical for power utilities to
have detailed models of their networks as well as advanced
simulation tools [5]. To ensure accuracy and efficiency in
this quantification, particularly in networks with thousands of
customers, it is important to develop these tools with pow-
erful modeling techniques and high performance computing
capabilities.
To make these simulation tools more attractive, flexible and
accessible, open source software packages should be used.
In this context, OpenDSS [6] is a powerful, open source
software developed by EPRI (USA) to model and simulate the
electrical behavior of the distribution network. Although it is
a script-based simulator with limited Graphical User Interface
(GUI), it can be driven from other platforms (e.g., Python
- a programming language to perform computing tasks [7])
through the Component Object Model (COM) server.
From the perspective of utilities, the development of these
simulation tools should consider their Geographical Informa-
tion System (GIS). This is critical as most of the data needed
to create network models (e.g., type and length of conductors,
type and location of customers, voltage levels, etc.) is stored
in their GIS platform, which typically cannot perform power
flows. However, the stored data can be efficiently translated to
the corresponding network models ready to be used in software
packages for power flow studies as detailed in [8].
The number of reported applications that combine the above
requirements is limited. A Monte Carlo simulation platform
implemented in Matlab R© for modeling and simulating only
LV networks considering 24h periods with a resolution of
one snapshot per second is presented in [9]. A real-time
simulator that adopts LabVIEW R© to allow hardware-in-the-
loop applications is discussed in [10]. A framework that helps
assessing the effects of communications on control actions is
introduced in [11]. Although these simulation tools can carry
out power flow studies, they usually require the adoption of a
proprietary software, they have been tested in small networks,
and/or they require the construction of the network model in
a platform that is unusual for engineers (new GUI).
This paper presents an OpenDSS-based simulation tool
driven by Python that is developed in an open source GIS
environment. The tool, which can be used in Spanish or En-
glish depending on the machine’s Operating System language,
takes advantage of the advanced modeling techniques and high
performance computing capabilities available in OpenDSS,
so detailed studies can be efficiently performed. Illustrations
on a real distribution network in Costa Rica with 10,000+
customers demonstrate its flexibility and efficiency to carry out
complex network studies (e.g., daily power flows), to display
results and to store them in comma separate value (csv) files.
This paper is organized as follows. Section II introduces
the open-source software used to create the simulation tool.
The OpenDSS-based distribution network analyzer in an open
source GIS environment is presented in Section III. Section IV
illustrates three different network studies that can be carried
out up to today. A discussion is presented in Section V and
conclusions are drawn in Section VI.
II. OPENDSS AND QGIS: OPEN SOURCE SOFTWARE
This section presents the open source software packages
used to develop the distribution network analyzer, which is
presented in section III as a plugin. These software packages
are selected to take advantage of their computing capabilities,
and also because they can be easily integrated, as detailed in
section II-C. This has the potential to increase the flexibility
and accessibility of the network analyzer.
A. OpenDSS
OpenDSS is an open source software package developed
by EPRI (USA) that can be used to simulate distribution
networks [6]. It contains detailed models of several network
components (e.g., lines, loads, generators, monitors, etc.). A
key aspect of this script-driven, frequency-domain simulation
tool is that it allows considering the time dimension (e.g., daily
simulations with different time step) - critical to quantifying
the impacts of variable sources and loads.
To create a network model, OpenDSS follows a sequence
of definitions [6], i.e., source, lines, transformers and loads.
However, given the script-written nature of this software, the
creation of large-scale networks must be done carefully. In
order to run this plugin, a network model needs to be added
by means of *.dss files. The latter can be created automatically
as reported in [8]. The database (other *.dss files) that contain
the characteristics and configuration of conductors is also used.
• WireData: Database that contains the characteristics (e.g.,
name, resistance, diameter and GMR) of each wire.
• Config Lines: Database that contains the geometry (e.g.,
number of conductors, phases and spacing) of each line.
• Loadshapes: Database that provides the electrical behav-
ior of a load along a given period.
• MVLines: File that details the connectivity of each MV
line/cable segment, geometry of line/cable and its length.
• Transformers: File that provides the information of all
transformers (e.g., losses, impedance, voltages, etc.).
• LVLines: File that details the connectivity of each LV
line/cable segment, geometry of line/cable and its length.
• ServicesLV: File that details the connectivity of each
service cable, its linecode and length.
• Loads: File that indicates load location, type, nominal
voltage and power factor and the associated loadshape.
OpenDSS is known to have a limited GUI. Although this
can be overcome using the COM server, which allows users
to drive OpenDSS from other platforms (e.g., Python), the
development of a GUI that will enable engineers to perform
advanced network studies in a more friendly way is critical.
This, in turn, is expected to further enhance the flexibility and
acceptability of this powerful software package.
B. Quantum GIS
Many power utilities use GISs to store, read, edit and
analyze data of the distribution network (referenced by a
spatial coordinate system [12]). A GIS platform cannot run
itself power flows. However, the GISs of utilities usually
contain the data needed to create detailed network models [8].
COM COM
Figure 1. Interfacing OpenDSS with QGIS via Python
One of the most popular open source GIS software package
is Quantum GIS (QGIS), under the GNU-GPL license. A key
feature of QGIS is that it allows the development of Python-
written plugins for extra data processing and analysis not
available in the existing software [13].
From the perspective of power utilities, the development
of Python-written plugins to integrate OpenDSS within their
GIS environment (for instance to create a user-friendly GUI
for OpenDSS) can be significantly beneficial as this allows
them to study their distribution networks, while performing
the corresponding analyzes within their database.
C. Integration of OpenDSS and QGIS
A plugin that drives OpenDSS from QGIS via the Python
console available in the latter has been developed to integrate
these software packages. This, in turn, results in a GUI that
is introduced below. To achieve this, the library comtypes.py
needs to be installed using the package management system
pip. This integration is illustrated in Fig. 1.
III. OPENDSS-BASED DISTRIBUTION NETWORK
ANALYZER IN QGIS ENVIRONMENT
The plugin (i.e., the OpenDSS-based distribution network
analyzer in QGIS environment) presented in this section is
developed in combination with the work discussed in [8].
These two works together represent part of the efforts made
at the University of Costa Rica to provide distribution com-
panies with the necessary simulation tools, so they can easily
understand the state of their networks, as well as to assess the
benefits and/or impacts of adopting low carbon technologies.
A. Plugin Architecture
The OpenDSS-based distribution network analyzer in QGIS
environment (named hereafter QGIS2RunOpenDSS) can be
used to carry out, so far, three different types of net-
work studies: snapshot, daily and harmonic power flows.
QGIS2RunOpenDSS can be used in Spanish or English. The
selection of the language can be done automatically, and
it is based on the machine’s Operating System language.
Translation has been achieved using QT Linguists.
Fig. 2 shows the architecture of QGIS2RunOpenDSS. A
user-friendly GUI developed in Qt Designer (formally pre-
sented below) is used to introduce key information of the
distribution network (e.g., frequency and voltage) as well as
to select the type of network study to be carried out. The
information is then read so as to upload the network model
and input data. Depending on the selected network study, the
corresponding workflow (i.e., function that contains the logic
of each network study) is performed. QGIS2RunOpenDSS
Tr
an
sf
or
m
er
C
irc
ui
t
S
tu
dy
00:00 03:00 06:00 09:00 12:00 15:00 18:00 21:00 00:00
3500
4000
4500
5000
5500
6000
6500
7000
7500
8000
8500
Results
Workflow
Network2modelUser2interface
Figure 2. Architecture of QGIS2RunOpenDSS
finally carries out the studies, displays and stores the simu-
lation results in a user-predefined folder. QGIS2RunOpenDSS
enables users to carry out these studies in a stochastic approach
(i.e., Monte Carlo simulations).
B. Graphical User Interface (GUI): Input Requirements
To perform network studies using QGIS2RunOpenDSS, the
GUI shown in Fig. 3 needs to be completed with some
electrical parameters of the circuit, which are usually known
by engineers. More specifically, users need to:
• Select the network to be assessed.
• Introduce basic information of the circuit and main trans-
former (e.g., voltage, frequency, phase angle and short
circuit currents). Default values extracted from the GIS
data are used otherwise.
• Select the demand curve monitored in the circuit (from
historical data).
• Select the type of study to be carried out (e.g., snapshot,
daily and/or harmonic power flows).
• Select the output folder where results will be saved.
• Define the number of simulations (if a Monte Carlo
approach is to be adopted).
C. Network Model
When the GUI is filled with the required informa-
tion, QGIS2RunOpenDSS imports the corresponding network
model. This is selected from a previously created database,
which has been produced here using the tool developed in [8].
QGIS2RunOpenDSS uses the input information to create
a master file. Based on architecture of OpenDSS, this file
contains the circuit definition (e.g., voltage, frequency and
main transformer) and redirects the files previously created
(WireData, Config Lines, Loadshapes, Lines, Transformers
and Loads). For instance, if the distribution network shown
in Fig. 4, which contains high voltage (HV), MV, and LV
networks, is converted to OpenDSS files, the master file for
this circuit will contain the information shown in Fig. 5.
QGIS2RunOpenDSS then reads the input data, which corre-
sponds in this paper to a pool of daily, real domestic load
profiles (15-min resolution, i.e., 96 points). This pool was
created using information from the utilities. In future studies,
data for industrial, commercial and electric vehicle [14] loads,
as well as photovoltaic generation can be used to produce
different load/generation pools. To illustrate the input data
used here, Fig. 6 shows four domestic load profile examples
with different energy consumptions, as well as the average
demand of the pool. Note that the average peak demand (circa
0.60 kW) occurs approximately at 6:00 pm.
D. Workflows for Network Studies
This section provides some details of the routines imple-
mented in Python to perform network studies. The routines
are based on OpenDSS architecture. The results to be dis-
played and stored are arbitrarily selected; however, this can
be changed as per utility’s needs. As discussed in Section V,
a plugin to visualize the results in QGIS is under development.
1) Snapshot Power Flow: QGIS2RunOpenDSS can be used
to carry out instantaneous power flows. This assessment is
useful to understand the network conditions during a specific
instant of the day (e.g., peak demand and minimum load). This
approach, however, might result in under or overestimation of
network impacts, particularly when considering the variability
of renewable energy sources (e.g., photovoltaic systems) and
loads (e.g., electric vehicles).
To solve snapshot power flows, the user also needs to select
the corresponding snapshot box (see Fig. 3) and define the date
(dd/mm/yyyy) and time (hh:mm) of the simulation, which are
used in a load allocation algorithm implemented in the plugin.
The load allocation algorithm allows matching the aggregated
demand of customers and losses with the monitored demand
of the whole circuit.
QGIS2RunOpenDSS allows assessing any instant of the
day. Due to the nature of the input data (15-min resolution),
snapshot simulations can be performed every quarter of an
hour (i.e., 0, 15, 30 and 45 min). This can be adapted so as
to consider other input data resolution. If the user sets a time
that is different to the resolution of the simulation, the plugin
finds the nearest available simulation point and the user is
notified of the change. For instance, if the user defines 18:05,
the simulation will actually be executed for 18:00.
Simulation results are displayed and saved in *.csv files in
the folder predefined by the user (a default folder is defined
otherwise). In this folder, a sub-folder named snapshot (i.e., in
\user defined folder\snapshot\) is created to store the results.
2) Daily Power Flows: QGIS2RunOpenDSS can also per-
form time-series daily power flows. This assessment is crucial
to quantify the impacts of variable sources and loads. To solve
daily power flows, the user needs to define the basic infor-
mation of the circuit, as well as to select the corresponding
daily box (see Fig. 3). The date (dd/mm/yyyy) of the year to
be assessed must also be defined, so as to perform the load
Figure 3. Graphical User Interface of QGIS2RunOpenDSS
HV MV
SOURCEBUS
BUSMV5
BUSMV4
BUSMV2
BUSMV1
BUSMV3
BUSMV6
BUSLV1
BUSLV2
BUSLV3
BUSLV4
BUSLV5
BUSLV6
LVMV
Figure 4. Example distribution network
clear
New Circuit.Test_circuit
set defaultbasefrequency=60
Edit Vsource.Source BasekV=___ pu=___ angle=___ 
   ~ frequency=___ phases=___ MVAsc3=___ MVAsc1=___
New transformer.HVMV phases=___ windings=___
   ~ buses=[sourcebus,BUSMV1] conns=[wye,wye]
   ~ kvs=[___,___] kvas=[___,___]
   ~ xhl=___ maxtap=___ mintap=___
redirect WireDataAAAC.dss
redirect WireDataAAC.dss
redirect WireDataACSR.dss
redirect WireDataCU.dss
redirect Config_Lines.dss
redirect LoadShapes.dss
redirect LinesLV.dss
redirect LinesMV.dss
redirect LoadsLV.dss
redirect ServicesLV.dss
redirect Transformers.dss
Figure 5. Sentences for defining a circuit in the plugin
0h 2h 4h 6h 8h 10h 12h 14h 16h 18h 20h 22h 24h
Time
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
A
ct
iv
e
 P
o
w
e
r,
 k
W
125 kWh/month
136 kWh/month
186 kWh/month
267 kWh/month
Average
Figure 6. Example of domestic load profiles
allocation algorithm. The load allocation is carried out here
simulation point by simulation point (i.e., every 15 min). Once
the daily power flow is solved, the results are displayed. The
plugin also stores the results in *.csv files in the user-defined
folder or default path (i.e., in \user defined folder\daily\).
3) Harmonic Power Flows: QGIS2RunOpenDSS can also
perform harmonic power flows. This study is crucial to assess
the quality of the service as well as to understand the impact of
harmonic distortion produced by loads and sources on distribu-
tion networks. This also allows determining the propagation of
current components of frequency other than the fundamental
and the resultant distortion on the voltage waveform. This
assessment is today even more important due to the increasing
adoption of new technologies with converters.
To solve harmonic power flows, the user needs to define
the basic information of the circuit, to select the harmonic
analysis box (see Fig. 3), to specify the number of harmonics
to be assessed and to define the spectrum of each load and
Substation HV/MV
LV overhead line
LV underground cable
MV overhead line
MV underground cable
Legend
Figure 7. Real Costa Rican circuit used to illustrate QGIS2RunOpenDSS
source. As before, the simulation results are displayed and the
plugin stores the results in *.csv files in the user-defined folder
or default path (i.e., in \user defined folder\harmonics\).
IV. ILLUSTRATION OF QGIS2RunOpenDSS
This section illustrates the results obtained using
QGIS2RunOpenDSS for each type of study. To do so,
the real distribution network shown in Fig. 7 is used. This
network is available in the GIS database of one of the largest
power utilities in Costa Rica. This network is operated
radially, and it consists of several transformers, MV segments
and LV segments supplying 13,324 customers. The OpenDSS
model of this network has 2,194 MV buses and 22,490
LV buses. Times reported below are based on a PC Intel R©
Core i5, 4GB RAM, 32 bits.
1) Snapshot Power Flows: Fig. 8 shows the normalized
frequency of occurrence of voltages at all 13,324 customer
connection points for an arbitrarily selected day at 6:00 pm
(peak time in Costa Rica). A snapshot simulation takes on
average 2.3 s. This includes the time needed in the load
allocation algorithm. QGIS2RunOpenDSS can also display the
current and voltages through monitored transformers and/or
lines. This, however, is not presented in this work. It must
be mentioned that the above analysis can be performed in a
Monte Carlo approach, and QGIS2RunOpenDSS will quantify
the average results with one standard deviation.
2) Daily Power Flows: Fig. 9 shows the active power
demand monitored in the actual circuit through the main
transformer, as well as the active power demand simu-
lated using QGIS2RunOpenDSS through the same asset. The
daily simulation (in a 15-min resolution) takes on average
50 s, including the runtime of the load allocation algorithm.
QGIS2RunOpenDSS allows users to export the simulation
results of any element (transformer, line and/or cable) in the
1.13 0.26 1.43
4.51
18.18
73.76
0.69 0.00 0.00 0.00 0.00 0.00 0.04
0.94 0.96 0.98 1.00 1.02 1.04 1.06
Voltage (p.u.)
0
20
40
60
80
N
or
m
al
iz
ed
 F
re
qu
en
cy
 (%
)
Figure 8. Normalized frequency of customer voltages in snapshot power flow
0h 2h 4h 6h 8h 10h 12h 14h 16h 18h 20h 22h 24h
Time of day
6
8
10
12
14
16
18
Ac
tiv
e 
Po
w
er
(M
W
)
Real demand
Simulated demand
Figure 9. Demand through the main transformer in a daily power flow
circuit. For this, the user only needs to select the corresponding
monitor from the pull-down list of monitors.
3) Harmonic Power Flows: Fig. 10 shows the distribution
of harmonics spectrum calculated at the main transformer
that supplies the circuit shown in Fig. 7. This distribution
has been found for an arbitrary day during peak time (i.e.,
6:00 pm). A harmonic power flow takes on average 20 s for
25 harmonics. It can be noted that QGIS2RunOpenDSS can
1 3 5 7 9 11 13 15 17 19 21 23 25
Harmonic number
0
10
20
30
40
50
R
M
S 
Cu
rre
nt
 (A
)
Phase A
Phase B
Phase C
Neutral
Figure 10. Distribution of harmonics during peak time
Table I
TOTAL HARMONIC DISTORTION FOR VOLTAGE AND CURRENT
THDV THDI
Phase A 0.47 19.18
Phase B 0.25 18.82
Phase C 0.35 21.09
calculated harmonics in each phase as well as in the neutral.
QGIS2RunOpenDSS also calculates the total harmonic dis-
tortion (THD) for both voltages (THDV) and currents (THDI).
To illustrate this, Table I shows THDV and THDI, per
phase, monitored at the main transformer of the circuit.
QGIS2RunOpenDSS is also capable of determining the Total
Demand Distortion, though this is not shown in this work.
V. DISCUSSION
QGIS2RunOpenDSS can perform, so far, snapshot, daily and
harmonic power flows. However, additional network analyzes
can be introduced so as to provide a wider range of possibil-
ities to understand the electrical behavior of the distribution
network. Indeed, the University of Costa Rica is working to
provide in the near future the possibility to carry out yearly
power flows, short circuit analyzes and network losses quan-
tification. In addition, although QGIS2RunOpenDSS allows
storing results, a new tool that enables engineers to visualize
the results in GIS environment is currently under development.
Finally, it must be mentioned that this simulation tool will
be used by the eight distribution companies in Costa Rica to
assess the future penetration of low carbon technologies in
their distribution networks.
VI. CONCLUSION
This paper has presented an OpenDSS-based simulation
tool in an open source GIS environment (i.e., QGIS) that
enables engineers to perform different network studies ranging
from snapshot to harmonic power flows. QGIS2RunOpenDSS
enables power utilities to rapidly and adequately assess their
distribution networks. QGIS2RunOpenDSS requires limited
input information to carry out these complex studies, thus the
daily network analyzes performed by engineers is facilitated.
The creation of this simulation tool not only allows power
utilities assessing their existing networks, but it also enables
them to understand the hosting capacity of their circuits, and to
study the impacts and future interactions of new technologies
with the distribution network. Moreover, it is likely that the
tools being developed at the University of Costa Rica will also
help quantifying the benefits of future Smart Grid schemes
(e.g., demand side management programs).
The performance of the simulation tool has been demon-
strated on a real circuit in Costa Rica with more than 10,000
customers. The illustrations presented here have highlighted
the efficiency of the tool to perform network studies, display
and store the corresponding results.
VII. ACKNOWLEDGEMENTS
The authors would like to thank Mr. Rau´l Ferna´ndez and
Mr. Guido Godı´nez, affiliated to CNFL, for providing the GIS
database of the analyzed circuit.
REFERENCES
[1] IEA, “Energy Technology Perspectives,” International Energy Agency,
Tech. Rep., 2015. [Online]. Available: http://www.iea.org/etp/etp2015/
[2] IPCC, “Climate change 2014: Impacts, adaptation, and vulnerability,”
Intergovernamental Panel on Climate Change, Tech. Rep., 2015.
[3] A. Navarro-Espinosa and L. F. Ochoa, “Probabilistic impact assessment
of low carbon technologies in LV distribution systems,” IEEE Trans.
Pow. Syst., vol. in Press, pp. 1–12, 2015.
[4] J. Quiros-Tortos, L. F. Ochoa, S. Alnaser, and T. Butler, “Control of EV
charging points for thermal and voltage management of LV networks,”
IEEE Trans. Pow. Syst., vol. in press, pp. 1–11, 2015.
[5] G. Shirek, B. A. Lassiter, W. Carr, and W. H. Kersting, “Modeling
secondary services in engineering and mapping,” IEEE Transactions on
Industry Applications, vol. 48, no. 1, pp. 254–262, Feb. 2012.
[6] R. Dugan and T. E. McDermott, “An open source platform for collab-
orating on smart grid research,” in IEEE PES General Meeting, 2011.
[7] Python Software Foundation, Python Language Reference, version 3.5.
[Online]. Available: http://www.python.org
[8] P. Quesada, A. Arguello, J. Quiros-Tortos, and G. Valverde, “Distribution
network model builder for OpenDSS in open source GIS software,” in
IEEE PES Transmission and Distribution Latin America, 2016.
[9] R. Torquato, Q. Shi, W. Xu, and W. Freitas, “A monte carlo simulation
platform for studying low voltage residential networks,” IEEE Trans.
Smart Grid, vol. 5, no. 6, pp. 2766–2776, Nov. 2014.
[10] D. Montenegro, M. Hernandez, and G. Ramos, “Real time OpenDSS
framework for distribution systems simulation and analysis,” in IEEE
PES Transmission and Distribution Latin America, 2012.
[11] A. Awad, P. Bazan, and R. German, “SGsim: A simulation framework
for smart grid applications,” in IEEE International Energycon, 2014.
[12] P. Bolstad, GIS Fundementals, a First Text on Geographical Information
Systems, 4th ed. Minnesota EEUU: Eider Press, 2002.
[13] QGIS, PyQGIS developer cookbook, February 2016. [Online]. Available:
http://docs.qgis.org
[14] J. Quiros-Tortos, L. F. Ochoa, and B. Lees, “A statistical analysis of
EV charging behavior in the UK,” in IEEE Innovative Smart Grid
Technologies - Latin America, 2015.