Modeling and Simulation of Distribution Networks with High Renewable Penetration in Open-Source Software: QGIS and OpenDSS

Abstract

There are important challenges in modeling large electrical distribution circuits, especially with the presence of distributed renewable generation. Constructing simulations to assess the effect of the penetration of distributed generation on electrical distribution networks has become of great importance for Distribution Network Operators (DNOs). This paper proposes a simulation strategy based on open-source platforms and the integration of scripting tools for the rapid modeling of large-scale electrical distribution circuits with distributed renewable generation. The implementation is based on the adaptation of a tool called QGIS2OpenDSS, which creates OpenDSS distribution network models directly from an open-source geographic information system, QGIS. The plugin’s capabilities are demonstrated using a real distribution feeder with more than 60% penetration of renewable generation based on photovoltaic systems. These simulations are carried out using real data from a circuit provided by a DNO in the Dominican Republic, which is used to demonstrate how this approach provides a more accessible and flexible way to simulate and assess the effect of Distributed Energy Resources (DERs) in medium voltage (MV) and low voltage (LV) networks, enabling utilities to evaluate system performance and identify potential issues. The integration of this open-source tool within the DNO software stack enables users to apply it according to specific project needs, enhancing their capability to analyze and manage high DER penetration levels, aiding in better planning, operation, and decision-making processes related to renewable energy projects.

1. Introduction

With the intention of effectively performing analysis on the impact of Distributed Energy Resources (DERs) in the medium voltage (MV) and low voltage (LV) networks, utilities must first have in place the data necessary to build detailed models of the Distribution Network (DN) components [1], as well as advanced simulation tools that allow for comprehensive analysis to evaluate system performance and identify potential issues [2].

The fact that DERs are geographically dispersed, and that their engineering treatment, i.e., energy planning, stationary or dynamic grid analysis and protection, mainly depend on some geographical qualities, makes Geographic Information System (GIS) environments ideal for their study [3]. Furthermore, GIS has become an important tool for utilities and city planners and has been embedded in their internal processes from asset and operations management to planning, design, and engineering. Various works on the scientific literature have focused on the integration of GIS environments for adding value to decision processes regarding the topology of electrical power systems and DER potential assessment [4,5,6,7].

To make the simulation of DER behavior more accessible and flexible, the use of open-source software tools such as OpenDSS^® (Open Distribution System Simulator) and open-source GIS software (QGIS) has been recently explored and used by utilities and research laboratories across the world [8,9,10,11,12].

However, simulation results with these tools highly depend on the quality and availability of network data (location and capacity of distribution transformers, capacitors, and DERs, as well as type, material, size, and length of conductors, etc.) [13], and the necessary work to translate (clean, fix, aggregate or remove incorrect, corrupted, incorrectly formatted, duplicated, or incomplete) data within the GIS network to the respective power flow. Simulation tool use is not only arduous, but often requires third party commercial software packages [14] that need licensing, and that do not allow access to the source code to change or add any functionalities and new algorithms from users.

Several works in the scientific literature present various approaches for the manipulation of the GIS database to generate automatically network models. The authors in [15,16,17] report on the construction of network models using MATLAB^® (2022b) for the extraction, processing, or network modeling. However, a common feature and liability of these studies is the fact that recurring modifications in the GIS databases require new interactions of file exchanges between software packages, a repetitive and time-consuming task. It is also often the case that studies of photovoltaic (PV) penetration are carried out using synthetic data of the DER characteristics [18,19,20,21,22] with simplified or generic networks; this reduces the accuracy on capturing specific regional or site-specific factors and limits the validation opportunities for the models and the correct representation of the real distribution circuit behavior [23].

This paper reports on the implementation and adaptation of a software plugin called QGIS2OpenDSS, developed by researchers at the Electric Power & Energy Research Laboratory (EPERLab) of the University of Costa Rica (UCR) [24]. This tool creates the OpenDSS network model directly from an open-source geographic information system (QGIS) using PyQGIS, QGIS library for executing Python code [25]. Although limited by its Graphical User Interface (GUI), OpenDSS has been used in several studies, harnessing the functionality of plugins that are capable of automatically scripting the required connection matrices [26]. Nevertheless, exploring the use of the QGIS2OpenDSS plugin in a real circuit with real data on DER penetration has not been widely investigated. This paper contribution focuses on adapting and applying the proposed simulation scheme to a real study case of a DNO. The main contributions are the following:

• Direct Integration of Real Data: This study implements the QGIS2OpenDSS plugin with real-world data from a georeferenced high PV penetration network, unlike studies which relied on synthetic data or simplified models. This ensures more accurate and reliable simulation results.
• Detailed Explanation on the procedure required for the use of the QGIS2OpenDSS plugin: The original publications presenting this plugin [8,27] do not discuss in detail the necessary steps for using the plugin in different regions. By addressing this gap, this work contributes to extending the usage of the plugin and documents the required adaptations and modifications for its implementation in other locations. Furthermore, guidelines for extracting the required data for the plugin from the DNO data are presented. This effort enhances the utility of the plugin and supports its broader applicability.
• Automated Data Processing: The development of an automated data processing algorithm allows for the conversion of legacy GIS data from the DNO into the format needed for QGIS2OpenDSS, reducing the time and effort required for manual data handling and minimizing the risk of errors associated with repetitive data exchanges.
• Open-Source and Customizable Tool: The use of open-source tools like QGIS and OpenDSS eliminates dependency on costly, proprietary software, and allows for customization. This enables users to modify the tools according to specific project needs and enhances flexibility in performing various analyses.
• Practical Application: By applying the developed tool in the context of the Dominican Republic’s distribution network, the study provides practical insights and validates the effectiveness of the tool in a real-world scenario, potentially accelerating DER integration and supporting the local energy market’s growth.

The integration of this tool in the software stack of the DNOs in the Dominican Republic is an important contribution. Recent studies on the issues regarding the development of DERs in the island have identified that the lack of access and capacity on specialized electrical system analysis software for DER interconnection procedures creates unnecessary delays on the approval of projects and the further development of the market due to the current established policies [28].

This paper is organized as follows: Section 2 includes a summary of QGIS2OpenDSS and the software stack behind it, as well as the required list of attributes for each DN element to build its model in OpenDSS and the adaptations made to the plugin for the case study. In addition, Section 3 explains the Distribution Network Data Extraction Process from a real DN feeder (VOLG101) with high renewable penetration (>60% PV) in Santiago de los Caballeros, Dominican Republic, with 7428 customers, while a demonstration of the plugin capabilities is shown in Section 4. Finally, Section 5 illustrates the type of analysis that can be carried out to study the impact of PV generation, and future work and conclusions are drawn in Section 6.

2. Materials and Methods

To perform effectively an analysis on the impact of Distributed Energy Resources (DERs) in the medium voltage (MV) and low voltage (LV) networks, utilities must first have in place the data necessary to build detailed models of the Distribution Network (DN) components, as well as advanced simulation tools that allow for comprehensive analysis to evaluate system performance and identify potential issues. This section presents a description of the open-source GIS software package in which the plugin was built (QGIS) and the power engineering software for which the models are created (OpenDSS), as well as a summary of the most important aspects of QGIS2OpenDSS plugin architecture and functionalities.

2.1. QGIS

Geographical Information Systems (GISs) are a powerful tool which can be defined as an integrated set of data, hardware, software, and processes designed as a computer system for gathering, managing, mapping, and analyzing spatial data [29]. GISs are now being widely used by distribution energy utilities to manage and store information about the characteristics of their assets, which in turn allows power engineers to use this data (overhead lines distance, conductor type, transformer data, customer location, and consumption, etc.) to create detailed network models.

QGIS is an Open-Source Geographic Information System (GIS) licensed under the GNU General Public License, and part of the Open-Source Geospatial Foundation (OSGeo), which allows developers to create plugins to add additional features that are not part of the original software stack [30].

2.2. OpenDSS

OpenDSS is a comprehensive electrical power system simulation tool primarily for electric utility power distribution systems. It is an open-source software package that supports nearly all frequency domain (RMS sinusoidal steady-state) analyses commonly performed on electric utility power distribution systems by the DNO [31].

A key aspect of this script-driven simulation tool is that it works with numerous solution modes including the standard single Snapshot mode, Daily mode, Duty Cycle Mode, Monte Carlo mode, and several modes where the load varies as a function of time, which is critical to quantifying the impacts of variable sources and loads [32]. OpenDSS differs from the traditional radial circuit solvers because it solves networked (meshed) distribution systems as easily as radial systems.

Figure 1 shows the internal architecture of the software and the sequence of definitions, i.e., power delivery, conversion, and control elements that contain the characteristics and configuration.

2.3. QGIS2OPENDSS Plugin

QGIS2OpenDSS extracts and processes GIS data to generate the OpenDSS files (*.dss) required to run the simulator. The plugin contains a Graphical User Interface (GUI) that facilitates the translation of data, that comes in the form of *.shp files, which contain the information, characteristics, and attributes of object classes (LV/MV lines, loads, transformers, etc.) (See Figure 2).

In the GUI, the user would select the layer of DN elements to be modeled in OpenDSS, each layer has attributes that are classified as mandatory (essential) and optional (these attributes can be extracted indirectly from the files generated by the GIS departments of the electricity distribution companies).

Relevant aspects of the plugin´s inner workings such as the connectivity of elements (done using Python Library Kd-trees [33]) and the load profile allocation algorithm (where a typical demand profile is assigned to each load) are out of the scope of this paper; an interested reader can refer to references [34].

2.3.1. Data Requirements in QGIS2OpenDSS

The plugin was originally designed to work with the CRTM05-WGS84 (Costa Rica) projection, but users can modify this if the projection coordinates are given in meters. This is important because the object location is a common attribute among all DN elements. For our case study UTM Zone 31-WGS84 was selected. Here is the list of some of the required DN elements to build a DN model, and an example of how the naming convention for the attributes works on the plugin for the overhead MV/LV lines and for PV DERs.

• Overhead MV/LV Lines

Table 1. Overhead MV/LV attributes.

Required Attributes	Optional Attributes
NEUTMAT	LENGTH
NEUTSIZ	LENUNIT
PHASEMAT	X1
PHASESIZ	Y1
LINEGEO	X2
PHASEDESIG	Y2
NOMVOLT

presents the list of attributes in the shape files of overhead medium voltage (MV), low voltage (LV) lines. A description of each parameter is presented below.

• • NEUTMAT: Neutral conductor material (CU, AAC, AAAC, ACSR).
  • NEUTZIZ: Neutral conductor gauge—Can be specified in AWG, mm2, MCM.
  • PHASEMAT: Phase conductor material (CU, AAC, AAAC, ACSR).
  • PHASESIZ: Phase conductor gauge—Can be specified in AWG, mm2, MCM.
  • LINEGEO: Geometry of the line—These data characterize the geometry used in the conductors. Only one letter is used to indicate the type (H—Horizontal, B- Biphasic, V-Vertical, T-Triangular).
  • PHASEDESIG: Designation of the phases. The user can use letters or numbers as encoding.
  • NOMVOLT: Coding for nominal voltage. The user must select one of the predefined codes for the voltage selection.

• Small Scale DERs (PV)

Table 2 shows the list of mandatory and optional attributes in the shapefiles of the distributed generators. These generators take the number of phases and connections of the loads where they are installed. A description of each parameter is presented below.

• • TECH: Distributed generator type (PV, hydro, wind).
  • KVA: Generator installed power in kVA.
  • CURVE1: Irradiance curve file name for photovoltaic systems, must include the extension.
  • Curve2: File name of the temperature curve for photovoltaic systems.

Table 2. Distributed generators attributes.

Required Attributes	Optional Attributes
Tech	X1
KVA	Y1
PHASEMAT
Curve 1 (Irradiance)
Curve 2 (temperature)

Table 2. Distributed generators attributes.

Required Attributes	Optional Attributes
Tech	X1
KVA	Y1
PHASEMAT
Curve 1 (Irradiance)
Curve 2 (temperature)

The user will need to verify and translate the existing utility network data to match the naming convention and requirement of the plugin for the rest of the DN elements:

2.3.2. Erroneous Data

One of the key features of the plugin is that if any of the attributes listed are missing, the user will be alerted with an error message. It is possible that the attribute name in the plugin may not match the name in the table, leading to non-recognition. The plugin is also capable of the following:

• Detecting disconnected elements due to small coordinate displacements.
• Wrong phase designation: The plugin verifies that the elements being connected have the correct phase designation.
• Unknown transformer model capacity or nominal voltage:

These errors must be fixed in the GIS database to successfully run the plugin.

2.3.3. Adaptations and Updates to the Plugin for the Dominican Republic Case Study

• Line configuration library.

Several new line geometries were added to the Line Configuration Library (see Figure 3). Each Line geometry object specifies the number of conductors (nconds), the number of phases (nphases), the spacing configuration (spacing), the wire types (wires), and whether the conductor information should be reduced for modeling (Reduce = Y). As expected, the distribution network code and regulation from the Dominican Republic differ from Costa Rica, so configurations such as depicted in Figure 4 were added, as well as new line geometries for underground cables. Procedures similar to those shown in [35] follow.

• List of reactance and resistance for single phase three winding transformers.

In the Dominican Republic, 12.47 kV feeders represent almost 90% of all MV feeders. Series impedances for the transformers at this nominal voltage level were added to the TrafoOperations.py script. The plugin uses a look up table of typical series impedances based on the nominal voltage and capacity of transformers. If the value is not found in the table, it will trigger an error.

• Function related to single-phase L-N voltage and three-phase L-L voltage assignation.

The function called renameVoltage (see Figure 5) was modified in the TrafoOperations.py; this function takes two input parameters, MVCode and LVCode. The purpose of the function is to convert medium voltage (MV) and low voltage (LV) codes into their respective voltage values for both line-to-neutral (LN) and line-to-line (LL) configurations. The commented part indicates that the voltage values for the MVCode 210 are specific to the Dominican Republic. This suggests that the voltages 7.22 kV (line-to-neutral) and 12.5 kV (line-to-line) are standard for medium voltage in the Dominican Republic.

• Functions related to service connection to loads.

A segment of the plugin qgis2opendss.py that determines which line the load is connected to, to validate if the wire code accepts the indicated SERVICE type, was modified to accept cable type==“DPX”. In general, the complete operation sequence of the plugin is shows on the flowchart in Figure 6, the red boxes are the scripts that were modified and adapted for the Dominican Republic case study.

3. Case Study: Distribution Network Data Extraction Process

With the goal of implementing the OpenDSS-Based Distribution Network Builder in a real case study, a blueprint of the processes to obtain, clean, and transform the existing GIS data of a regional distribution energy company in the Dominican Republic (EDENORTE) was created (see Figure 7). Several data dictionaries, models, and workflows were created in QGIS to standardize and summarize the process.

The test feeder (VOLG 101) characteristics are shown in

Table 3. VOLG101 characteristics.

Circuit Alias	VOLG101
System voltage (kV):	12.47
Number of customers:	7428
Sub transmission voltage (kV):	69
Circuit MV lines (kM):	78
Circuit LV and service lines (kM):	145
Number of transformers:	578
Number of PV installations:	394
Reported technical losses (2021)	6.4%
Renewable penetration:	60%

.

QGIS Model Designer was used to create a model for data extraction and translation with a simple and easy-to-use graphical interface (See Figure 8). This was a key development because, while working on the GIS datasets, most of the analysis operations that were executed were not isolated but were part of a set of operations. A model was used such that the chain of processes could be executed in a single process or algorithm, saving time and effort, which is proportional to the volume of the model.

The algorithm facilitates the transformation of network elements such as transformers, and assigning values to key attributes based on specific conditions derived from existing tabular data. By systematically altering attributes such as phase designation, primary and secondary voltages, connection types, and power ratings, the algorithm streamlines the process of data preparation for subsequent use in the plugin.

Another model was used to extract MV/LV loads from the registered consumption in kilowatt-hours by clients (monthly energy consumption), which is the equivalent of the KWHMONTH attribute for the plugin.

For the irradiance data used in the simulation, physical measurements were taken from a sensor located in the selected circuit for the case study. The Pyranometer SI-485 provides measurements in the spectral range between 300–2900 nm, with a maximum irradiance measurement of 1600 W/m² with an output resolution of 1 W/m² [36].

Table 4. Statistical summary of the data used to model photovoltaic generation.

Hour	Average	Minimum	Maximum	50th Percentile	Standard Deviation	Selected Day
						28 March 2023
23:45	0.000	0.000	0.000	0.000	0.000	0
22:45	0.000	0.000	0.000	0.000	0.000	0
21:45	0.000	0.000	0.000	0.000	0.000	0
20:45	0.000	0.000	0.000	0.000	0.000	0
19:45	0.001	0.000	0.179	0.000	0.010	0
18:45	2.375	0.000	22.134	0.196	4.264	0.017884
17:45	29.752	0.179	144.252	22.156	26.983	41.68653
16:45	130.993	1.717	332.588	119.907	75.274	191.4456
15:45	289.866	1.717	532.806	292.279	139.361	378.2549
14:45	483.396	1.198	773.078	533.772	187.047	665.2928
13:45	653.514	1.169	909.129	718.881	218.493	825.4596
12:45	771.808	0.828	1033.303	840.484	219.235	1013.152
11:45	812.093	0.828	1052.026	890.686	223.063	1052.026
10:45	770.307	0.828	1025.436	840.943	209.990	1025.436
9:45	654.453	0.828	948.757	714.060	188.165	948.757
8:45	476.251	0.828	699.502	512.404	155.591	693.829
7:45	261.228	18.931	452.867	269.825	110.564	450.0188
6:45	71.061	1.676	203.709	65.157	51.991	203.7091
5:45	3.839	0.000	37.269	1.539	5.386	37.26907
4:45	0	0	0	0	0	0
3:45	0	0	0	0	0	0
2:45	0	0	0	0	0	0
1:45	0	0	0	0	0	0
0:45	0	0	0	0	0	0

shows the statistics of one year of measured data; the highest irradiance measurement registered was the selection criterium for choosing the irradiance curve for the proposed case study. This single day was selected to represent the most critical case of renewable penetration in the distribution circuit.

4. Demonstration

QGIS2OpenDSS File Creator

Once all issues regarding connectivity and naming conventions are cleared, the plugin takes the attributes of the GIS layers to generate the OpenDSS files. The plugin execution time averages 3 min approximately, on a PC INTEL^® Core I7, 8 GB RAM, 64 bits, running QGIS 3.18 Zürich. In this case study, the files generated on the directory by the plugin are shown in Figure 9.

The plugin also creates a list of buses for all the LV and MV elements in a *csv file, which can be used to plot the circuit and see power flow simulations in OpenDSS 9.4.1.1 and, which can be seen in Figure 10.

Each individual file created on the project directory is crucial for the execution of the power flow simulation on OpenDSS. The description of the function of these files is explained below:

• DG.dss: Includes the bus name, phases, nominal voltage, connection type, power rating, irradiance, and temperature. Figure 11 shows the typical irradiance curve in Santiago used to model the PV generation.

• MVLines.dss: Details the connectivity of each MVline/cable segment, geometry of line/cable, and its length.
• LVLines.dss: Details the connectivity of each LV line/cable segment, geometry of line/cable, and its length.
• MV/LVLoads.dss: Indicates load location, type, nominal voltage, and power factor and the associated load-shape.
• Substation.dss: Provides information on the source bus, phases, connection, windings, power rating.
• Transformers.dss: Provides information on all transformers (e.g., losses, impedance, voltages, etc.).

Furthermore, a set of supplementary libraries with information regarding load profiles, wire information, and the geometry configuration for the distribution network must be available. This is also produced by the plugin and organized in the following manner:

• Wiredata: Database that contains the characteristics (name, resistance, diameter, and GMR, etc.) of each wire.
• ConfigLines: Database that contains the geometry (spacing, number of conductors, phases) of each line.
• Loadshapes: Provides the electrical behavior of a load along a given period.

In total, the plugin allocated 451 load profiles among 7528 customers. The circuit provided by the DNO is mostly residential, with only 10% of the customers being commercial. This is relevant when looking at the average load shape in Figure 12. Additionally, this form of residential load shape is consistent with profiles of residential circuits reported in the scientific literature, as shown in [37].

5. Results and Discussion

5.1. Simulation in OpenDSS

This section illustrates various studies that have been carried out to study the impact of photovoltaic (PV) generation in the proposed case study, showcasing the integration of OpenDSS and QGIS.

Figure 13 presents the GIS model of the circuit provided by EDENORTE. The blue circles represent loads, black triangles MV transformers, while yellow circles are the PV generators. This model is made of 10,206 LV and 2470 MV buses, 9990 LV line sections and wire services, 873 (3φ) MV, 4 (2φ) MV, and 1621 (1φ) line sections, and 579 MV transformers.

Once all OpenDSS files *.dss were generated as shown in Section 4, a snapshot power flow simulation was carried out to quantify the impacts of variable sources and loads in the voltage profiles of the feeder.

Figure 14 depicts line-to-neutral (L-N) voltage profiles in per unit (pu) over a distance (in kilometers) for two scenarios: (a) without photovoltaic (PV) generation (left graph) and (b) with PV generation (right graph). The X-axis (distance in km) represents the distance along the distribution line from the substation, while the Y-Axis (p.u. voltage) shows the voltage levels per unit (pu), where 1.0 pu represents the nominal or rated voltage. The different colors represent different voltage phases. The slope of the voltage profile is less steep in the right graph, suggesting that PV generation has a stabilizing effect on the voltage levels along the distribution line.

Figure 15 and Figure 16 present the variation of all MV/LV buses throughout the day. There is significant variability in the LV/MV bus voltages, however mostly within the 120 V to 130 V range, suggesting that the system is generally well-regulated. There are notable voltage drops in the early morning (6:00 a.m.) and evening (6:00 p.m.). These times correspond to periods of increasing demand and low PV contributions.

The effect of PV integration on the technical losses in the DN was also evaluated. For this purpose, a daily time series simulation was carried out under two conditions: no PV systems and the inclusion of the PV systems that were present in the DNO GIS database. Figure 17 shows the results of the feeder load profile, which is consistent with the residential nature of the circuit, as can be seen in Figure 10.

This time series simulation allows us to identify the critical hours on which to explore the effects of PV system penetration on technical losses. As can be seen, around 12:00 p.m. is the moment in the day on which the effect of the PV penetration causes important changes in the behavior of the main electrical feeder of the circuit under study. As stated in Section 4 of this paper, actual irradiance information was captured for the construction of this simulation results. From this information, and using the power flow results, it is possible to construct a heat map for the zones on which the PV penetration impact is greater in terms of percentage voltage rise, as shown in Figure 18. The y-axis represents the percentage voltage rise, indicating how much the voltage levels have increased due to the addition of PV generation. Each dot on the plot represents a measurement point, and the color coding visually differentiates the magnitude of the voltage rise.

On average, the buses with the highest percentage of voltage rise (16%) are those close to the highest concentration of PV while the lower voltage rise is found at the tails of the secondary circuits and where the PV concentration is the lowest.

Finally, to measure the effect of these changes on the behavior of the circuit, it is important to determine if there is a global reduction or increase in technical losses. As shown in Figure 19, the total circuit losses are reduced from 238.83 kW to 209.35 kW for an improvement of 12.34%. The summary of the power flow simulation is shown in Figure 20.

5.2. Discussion

As demonstrated, detailed studies on real geographically dispersed DERs can be easily carried out with the integration of GIS and advanced simulation tools such as OpenDSS. These tools facilitate the assessment, accurate representation, and analysis of complex distribution networks facing increasing integration of renewable energy sources, allowing for the evaluation of various technical aspects such as voltage regulation, power quality, and system stability. This creates the possibility of conducting studies of the Dominican Republics’ electrical grid with similar focus to the ones shown in [38,39,40].

For validation purposes, the obtained results were compared with real data provided by the DNO regarding the reported average technical losses for the year 2021 (as shown in Table 3. As can be seen in Figure 20, the absolute error between this value and the one reported by the DNO is approximately 9.5%. This implies that the topological arrangement and considerations that were taken in the modeling process are in line with the real behavior of the circuit. The differences between the reported technical losses value and the simulation results can be explained by the previous exposed modeling challenges regarding the DNO dataset. Thus, the methodology presented by this work could be implemented for the simulation of the actual DNO distribution circuit topology.

This subject holds utmost significance, particularly in countries like Dominican Republic, where present regulations require a supplementary study for new photovoltaic systems if the actual installed capacity in that circuits exceeds 15% of the circuit´s annual peak load [41]. The additional cost of such a study constitutes an obstacle to the installation of new photovoltaic systems, for small-scale photovoltaic systems where the cost is disproportionately high. In addition, it creates a backlog in the studies on the DNO side since it needs to process many studies through applications above the 15% limit.

Recent studies have demonstrated that the technical limit is often much higher than the proposed limit of 15% by the regulations in the Dominican Republic. Even extreme feeders with PV distribution at the end show at least 25% accommodation capacity with mitigation measures such as voltage setpoint optimization or reactive power control, and even 50% in urban feeders [42]. As the country seeks to increase its renewable energy generation, understanding the maximum penetration level of PV installations is crucial for ensuring the stability and reliability of the distribution grid, as well as to help in planning and implementing sustainable energy strategies, fostering the growth of renewable energy, and reducing dependence on fossil fuels.

However, from the author’s experience, the DNOs in the Dominican Republic lack access to tools to conduct these detailed studies; therefore an open-source repository with the processes and models to clean, fix, aggregate, or remove incorrect, corrupted, incorrectly formatted, duplicated, or incomplete data within the GIS network dataset was created to enable the data to be processed by the QGIS2OPENDSS plugin in QGIS, which is the key in carrying out these studies.

Through the validation process some errors were found in the network element database, requiring additional network analysis for correction. Namely, there are discrepancies in the association of loads to transformers, as well as energy consumption data in the DNO database. This database must undergo a thorough cleansing process to remove any erroneous information that may lead to the creation of impractical engineering network models. Correcting these errors is a complex task that demands an understanding of the construction of GIS network models and their appropriate adaptation for power engineering purposes.

Finally, to fully harness the capabilities of GIS software, forthcoming efforts will concentrate on the development of hosting capacity maps and investigations pertaining to the penetration of electric vehicles (EVs). Additionally, GIS-based tools will be devised to facilitate the mapping and identification of infrastructure weakness, such as trees located near overhead medium voltage (MV) lines, among other location-based hazards. For example, preliminary analysis conducted on the DN data reveals that the occurrence of tree limbs falling into overhead distribution lines constitutes a contributing factor to nearly 10% of power outages within certain distribution feeders.

6. Conclusions

Overall, this paper contributes to the growing body of research on modeling and simulation of distribution networks with high renewable penetration. The proposed framework utilizing QGIS and OpenDSS provides a valuable tool for network planning, analysis, and optimization in the context of increasing renewable energy integration. It offers a cost-effective and flexible solution, empowering stakeholders to make informed decisions and pave the way towards a sustainable and resilient energy future.

The power of the tool was demonstrated on a real circuit in Dominican Republic with high renewable penetration and with more than 7000 customers. The illustrative network studies highlight the convenience of the tool to perform effectively analysis on the impact of Distributed Energy Resources (DERs) in the medium voltage (MV) and low voltage (LV) networks.

The data obtained from the simulations is a key input for the DNO regarding the penetration of renewable energy inside the distribution circuit selected for the case study. Even though penetration levels above 50% were presented, the voltage profile and congestion level shown in the simulations results are within regulation thresholds. Moreover, positive effects when considering voltage profile at some buses were observed, improving the case for the inclusion of higher renewable penetration levels in the circuit.

Furthermore, the possibility of using this tool for the creation of co-simulation scenarios will be explored in future projects, with the aim of creating toolboxes for the development of digital twin technologies focused on the operation of distribution network and fault detection processes.

The implementation of the tool using real data is expected to bring new opportunities for capacity building within the DNO, as well as future collaborations from the remaining DNOs and other research groups within the DR.