1 Introduction

The installation of Distributed Generation (DG) in electric distribution systems has shown significant growth rates in Brazil, reaching approximately 23 GW by 2023 (ANEEL, 2023). This growth follows a global trend, considering that between 2019 and 2021, 167 GW of photovoltaic DG were installed worldwide (IEAspsInternational Energy Agency, 2022). However, as the demand for new connections increases, so do the restrictions and limitations for new DG to access the distribution networks. Thus, there is an increasing need for detailed local quantitative studies that enable the assessment of the impact produced by the operation of DG resources and identify areas with the potential for connecting new units.

Numerous efforts have been undertaken to assess the impacts of DG on distribution networks. A fundamental issue is the amount of injected power into a given feeder without breaching operational limits. Studies addressing this question are concerned with the concept of network hosting capacity. Some qualitative conclusions can be drawn from these studies, such as the fact that as the injected power from DGs increases, overvoltage issues start to emerge, as pointed out by Mulenga et al. (2020). According to Chathurangi et al. (2022), overvoltages stand as the primary technical challenge arising from photovoltaic DG. This type of generation accounts for over 98% of the installed capacity in Brazil (EPE, 2022). Regardless of solutions to address voltage-related problems stemming from high DG penetration, such as efforts made by Zhang et al. (2022) and Lachovicz et al. (2023), quantitative analyses lack feeder-specific information essential for calculating hosting capacity. Concurrently, contemporary high-power loads, such as Electric Vehicles (EVs), have the potential to strain feeders during periods of reduced availability of intermittent energy sources. This situation necessitates the development of strategies to effectively manage grid congestion, as discussed by Hennig et al. (2023).

In addition to voltage level impacts, other energy quality parameters are also influenced by the presence of DG in the distribution network. Gimenes et al. (2022) evaluate the impacts caused by harmonic distortion, considering varying levels of DG penetration in the system. Indeed, the harmonic pollution close to the connection points of DG units proves to be significant. In this sense, it is not feasible to forgo tailored studies for specific feeders under specific operational conditions. Approaches to mitigate harmonic pollution through smart inverters and passive filters have been addressed by Silva Júnior et al. (2023) and Bajaj and Singh (2022). Another important consideration involves the impact of DG on short-circuit levels within the network. The presence of DG significantly changes these levels, which presents a challenge regarding planning and operating traditional protection systems (Neiva et al., 2021).

Furthermore, network expansion planning is of vital importance for the economic sustainability of the distributor, as well as for tariff affordability. For this reason, research in this area considering the presence of DG has been carried out in recent years, such as those carried out by Fiorotti et al. (2019) and Das et al. (2023). The significance of these studies grows as investments in reconductoring, allocation of voltage regulators, and transformers can quickly become outdated due to the expansion of consumer-based generation. Even the business models of distributors and their relationship with customers can change (Valarezo et al., 2023; Honarmand et al., 2021). Hence, realistic simulations of network expansion and operation are imperative to ensure compliance with technical requirements in light of potential market and policy-driven transformations (Costa & Bonatto, 2023).

At this point, it is worth noting that many of these works, which investigate the impact of the presence of distributed energy resources (DERs) on different fronts, use test distribution systems, like the standard models from the IEEE PES Test Feeder (IEEE Power & Energy Society, 2024). These standards are of vital importance in the context of the field addressed in this work. However, due to the dynamic nature of electric distribution systems, it can be said that each feeder is unique, highlighting the necessity of customized feeder models for local quantitative studies, including network operation, expansion planning, and renewable integration in general. On the other hand, it’s worth noting that obtaining actual data from distribution networks is not a straightforward task. In the references Peppanen et al. (2016) and (Wang & Luan, 2014), the main goal is to obtain, or calibrate, feeders models from AMI (advanced metering infrastructure) data. While the methodologies have many strengths, they need physical infrastructures for measuring and communicating systems, which increases the implementation costs.

In this context, this paper proposes a new methodological framework for extracting and processing information from Brazilian public data to create up-to-date electric feeder models. These feeder models can be utilized in local quantitative studies by utilities, investors, and consumers. Different topologies and loads in different circuits do affect hosting capacity, considering voltage and harmonic issues, expansion planning, short-circuit analysis, and protection scheme planning. To validate the methodology, a set of structured simulations is conducted using an open and free platform integrated with computational tools to analyze the impacts of connecting new DG in two real distribution systems.

In addition to this introductory section, the remainder of this article is organized into five main sections. Section 2 provides a brief overview of the Brazilian regulatory framework, primarily focusing on aspects related to the provision and access to data in the distribution system. Section 3 presents details concerning the BDGD and the process of requesting access to this database. Section 4 outlines the steps comprising the proposed methodology for generating feeder models via BDGD. Illustrative case studies applying this methodology are analyzed in Sect. 5, where two actual feeders from distribution companies CEMIG (Minas Gerais Electric Power Company) and Neoenergia COELBA (Bahia State Electricity Company) are modeled and evaluated via OpenDSS simulation, considering integration of DG units. Finally, Sect. 6 presents the main findings and concluding remarks.

2 Regulatory Framework—An Overview

The distribution of electricity in Brazil operates under a concession/permission public service system, within a regulated market where companies or consortia hold federal concessions granted by the government to distribute electricity in specific regions. In this regulatory framework (Gucciardi Garcez, 2017), as defined by the Regulations for Electric Power Distribution Procedures in the National Electric System (PRODIST) (ANEEL, 2021), distribution companies are required to provide the necessary information to the Brazilian National Electric Energy Agency (ANEEL) for tariff adjustments and reviews, along with a technical and financial survey of all their equipment, networks, and customers. All data related to the utility are made available in a common database known as BDGD (Portuguese acronym for utilities’ geographic database). Access to this data is guaranteed by the Access to Information Law (Presidency of the Republic, 2011), and can be requested from ANEEL or downloaded directly from the open database provided by the agency. This data is provided and stored in the BDGD in a standardized manner.

Therefore, based on the BDGD, this article proposes a method capable of modeling, in an automated way, three-phased feeder models of the electrical energy distribution system to perform power flow simulations via OpenDSS (EPRI, 2023), which is a software recommended by ANEEL as a reliable tool on these purposes (ANEEL, 2014). With the simulation from these feeders, it is possible to carry out evaluations related to the impacts of DGs on the network, such as studies of the hosting capacity in the network, investigations related to the quality of the power supply, short-circuit analysis, and general planning activities, including the identification of potential areas for investment and integration of new units.

The Brazilian regulatory procedures, as addressed in this study, play a significant role in the context of internationalizing regulatory practices. While considering adaptation to local realities, sharing best practices facilitates the shortcut toward a secure regulatory framework that fosters both sustainable and economic development. Partnerships and cooperation signed between Brazil and the European Union reinforce recognition in relation to national regulatory structures (EU-Brazil Sector Dialogues, 2020). In this regard, ANEEL has recently earned international recognition from the Organization for Economic Co-operation and Development - OECD (2021)—for its initiatives in Brazil, a country known for having one of the largest interconnected electrical systems globally. This regulatory work must be an ongoing effort to keep pace with technological and societal transformations. The expansion of DG aligns with the decentralization trend that is prevalent in Brazil (Maestri & Andrade, 2022).

3 Utility Geographic Database-BDGD

As defined in Module 10 of ANEEL’s distribution procedures (ANEEL, 2021), all distribution companies in Brazil are required to provide the principal technical, economic, and legal information about their assets. To expedite the standardization, ANEEL established the Regulatory Geographic Information System (SIG-R in Portuguese) and the BDGD. These databases aim to structure the data for registering the topology of distribution networks and support regulatory activities, assist in periodic tariff revision processes, and aid in technical and financial oversight. This prompted distribution companies to conduct surveys of their assets, registering the various components of distribution networks along with their respective georeferenced coordinates (ANEEL, 2022).

The SIG-R comprises a compilation of systems and databases aggregated by ANEEL, which collectively enable the extraction of diverse information regarding the distribution system and its users. The essential composition of SIG-R is depicted in the diagram shown in Fig. 1. In addition to BDGD, the SIG-R includes other data that expand the scope of information analysis, such as (with Portuguese acronyms addendum):

  • ANEEL Data Dictionary—DDA;

  • Electric Sector Asset Control Manual—MCPSE;

  • Electric Sector Accounting Manual—MCSE;

  • Brazilian Institute of Geography and Statistics Database—IBGE;

  • National Classification of Economic Activities—CNAE.

In essence, the geographic model established for the BDGD simplifies the actual electrical system for a specified period, aiming to reflect both the asset situation and the pertinent technical and commercial information. The BDGD’s geographic model primarily encompasses: (i) the geometric layout of network segments spanning high, medium, and low voltages, situated between each network support structure; (ii) the geographic positioning of network support structures; (iii) the geographic location of users and equipment; and (iv) the delineation of substations and other characterization areas. This database seeks to depict information linked to the technical data of the distribution system, commercial particulars, and the physical and accounting aspects of the asset base.

The BDGD is broadly divided into two types of information sets categorized as geographical and non-geographical entities. Concerning the geographical entities that portray features and informational structures, the data is inherently linked to ground coordinates. The non-geographical entities record technical and asset information and, as a rule, are associated with the component units of geographical entities to represent their location and the aggregation of two or more pieces of equipment operating together. For example, the geographic entity of a medium voltage transformer needs to be related to a non-geographic entity detailing its characteristics. These characteristics are related to complementary reference information about network elements, such as the date of acquisition, the start of operation, and nominal values of voltage, power, and electrical parameters. This is crucial due to the dynamic nature of distribution circuits, loads, and distributed energy resources (DER). The complete list of descriptions for these entities is outlined in the BDGD instruction manual (ANEEL, 2021).

Fig. 1
figure 1

SIG-R structure

Full size image

The data comprising the BDGD are public, guaranteed by the Access to Information Law (Presidency of the Republic, 2011), and can be requested from ANEEL or downloaded directly from the open database provided by the agencyFootnote 1. The files have the “.gdb” extension and can be accessed through the free software QGISFootnote 2

4 Proposed Methodology

This article introduces an automated method for modeling authentic distribution network feeders using BDGD for simulation in the OpenDSS software, which allows unbalanced three-phased power flow calculations. These feeder models can be utilized in local quantitative studies by utilities, investors, and consumers, as different configurations of branches and loads significantly influence the assessment of the distribution network. Common studies that could benefit from the feeder models include DER integration and hosting capacity assessment, reconfiguration, expansion planning, short-circuit analysis, and protection schemes planning. The methodological steps are illustrated in the diagram presented in Fig. 2, and detailed elaboration is provided in the subsequent subsections.

4.1 Step 1: Access to the Database

The first step to be executed corresponds to accessing the SIG-R. This database consists of a compilation gathered by ANEEL, which enables obtaining various details about the distribution system and its customers. As described in Sect. 3, within SIG-R lies the BDGD, a database containing the technical information of all equipment and networks operating within the system, as well as consumption details of all customers connected to the distribution grid. Accessing this information can be requested through the "Fala.BR" system, provided by the Brazilian Office of the Comptroller General—CGU (2023).

Fig. 2
figure 2

Overview of basic steps of the proposed methodology

Full size image

4.2 Step 2: Selection, Filtering, and Data Export

From the extensive database, only a subset of information is utilized to compose the feeder circuit for the distribution system of interest. Therefore, Step 2 of the methodology involves selecting the relevant information. When accessing the data via QGIS software, specific layers of information are chosen for effective integration into the feeder modeling process:

  • UNTRD (Portuguese acronym for Distribution Transformer Unit): contains the locations of the installed distribution transformer units in the system;

  • UCBT (Portuguese acronym for Low Voltage Consumer Unit): contains the locations of low-voltage consumer units or connection points with consumption characteristics;

  • UCMT (Portuguese acronym for Medium Voltage Consumer Unit): contains the locations of medium-voltage consumer units or connection points with consumption characteristics;

  • SSDBT (Portuguese acronym for Segment of the Low Voltage Distribution System): contains the layouts of low-voltage distribution network segments between support structures;

  • SSDMT (Portuguese acronym for Segment of the Medium Voltage Distribution System): contains the layouts of medium-voltage distribution network segments between support structures;

  • UNSEMT (Portuguese acronym for Medium Voltage Disconnecting Unit): contains the locations of installed medium-voltage sectionalizing units in the distribution network;

  • RAMLIG (Portuguese acronym for Connection Branch): each record in this entity represents a connection branch of a consumer;

  • EQTRD (Portuguese acronym for Distribution Transformer Equipment): each record in this entity represents an installed distribution transformer equipment in the system and composes the UNTRD entity;

  • SEGCON (Portuguese acronym for Conductor Segment): each record in this entity represents a grouping of existing conductor types in the distribution system.

The mentioned layers, which are described in greater detail in the BDGD manual (ANEEL, 2022), provide the technical information for modeling the three-phased OpenDSS feeder models with their respective loads. However, this information encompasses all the existing distribution networks in the BDGD, which is a massive repository. Therefore, it is necessary to filter the data according to the required feeder. According to the instructions from the BDGD manual, all utility networks and equipment must be identified by a unique code for each element. This code is then referenced in all other related structures. Thus, the filtering process begins by selecting the feeder to be simulated, followed by a selection of elements connected to the network using the feeder’s identifying code. This last part is carried out through the filtering capabilities available in the QGIS software.

Finally, following the completion of the filtration process, the data can be exported to generate the files for OpenDSS simulation, as outlined in Step 3 of the methodology. Data exportation is conducted in files with the ’.csv’ extension, a procedure that is also facilitated through the QGIS software’s features.

4.3 Step 3: OpenDSS Files Generation

The third step of the proposed methodology concerns crafting a set of files for simulation using the accessed, selected, filtered, and exported data from the preceding stages. These files configure the circuitry of the designated feeder within the OpenDSS software. To accomplish this, Step 3 can be executed through a computational routine.

This article proposes generating 8 ".dss" files, which follow the structures defined in the OpenDSS software manualFootnote 3 These files are indicated in Step 3 of the illustrative diagram in Fig. 2, and they are itemized as follows:

  • “Transformers-LV": in this file, the technical information of all connected distribution transformers is presented;

  • “Loads-LV” and “Loads-MV”: they present the loads of the system connected in low and medium voltage, respectively;

  • “Lines-LV”, “Lines-MV”, and “Branches-LV”: in these files, respectively, the information on the low and medium voltage networks and the connection branch between the load and low-voltage network are presented. Although the information is organized into separate files for clarity and compartmentalization, their structures are identical;

  • “Switches-MV”: in this file, information about the medium voltage circuit switches is defined;

  • “LineCode”: this file contains the physical characteristics of the conductors used in the network. Its utilization facilitates the organized and systematic construction of line files by aggregating all relevant information. The decision to use a separate file is influenced by the existence of a specific layer in the BDGD dedicated to this information.

Fig. 3
figure 3

Example of converting information to construct a command line for the ’transformers-LV’ file type

Full size image

As an illustrative example of how to use the data acquired in Step 2 to generate the OpenDSS files in Step 3, Fig. 3 presents a command line for OpenDSS related to the ’Transformers-LV’ file type. It’s important to note that certain data elements are used directly, while others require a conversion process, which is done, in this work, in the methodological framework by coding in Python language. The complementary OpenDSS code follows a similar approach.

Finally, to enable the simulation of the designated feeder using the OpenDSS software, in addition to the 8 files from BDGD, another set of 4 auxiliary files must also be created:

  • “Circuit”: In this file, the system’s primary bus is represented, considered by OpenDSS as a Thevenin equivalent of the entire circuit upstream of the distribution substation. This element includes the circuit’s base voltage (in kV), operating voltage (in per unit);

  • “BaseVoltages”: in this file, base voltages are declared for all voltage levels within the circuit;

  • “LoadCurves-LV” and “LoadCurves-MV”: in these files, the daily power measurements (provided by the utilities for the regulatory tariff revision process) for low and medium voltage are given, respectively. The values are presented in pu and normalized based on a declared daily maximum. A total of 24 measurements are provided at hourly intervals.

4.4 Step 4: DG Grid Impact Assessment

Within the scope of DG, the process of assessing its impact can be made more efficient and dynamic through automated and interactive methods. Thus, the final step of the proposed methodology is the DG Automated Interactive Impact Assessment (AIIA).

In the AIIA, the following substeps describe the main aspects of the comprehensive approach to DG impact assessment:

  • Define the pre-operational point: Before delving into the impact assessment, it is crucial to establish the pre-operational point (POP). The POP, also known as base-case, is the baseline condition against which the effects of DG will be measured.

  • Define the generation dispatch capacity factor/load scenarios: Specify the generation dispatch capacity factor and consider various load scenarios. Understanding how the DG system responds to different levels of demand and capacity factors is essential for a comprehensive impact assessment.

  • Define the connection point for new DG: Identify and define the point of connection for new DG installations. This could be at the substation bus, medium voltage (MV) feeder, or low voltage (LV) grid. The choice of connection point significantly influences the overall impact on the distribution network.

  • Define the new DG power capacity: The additional power capacity of the new DG source can be set as a target value (or increased in a predefined step during the simulation). For a given operation point, in an impact assessment, the critical condition usually is the maximum power output the DG system can provide to the network.

  • Increase power injection from DG (current injection scenarios): Systematically increase power injection into the network, considering various current injection scenarios. This step involves progressively assessing the impact of higher power inputs until a regulated electrical measurement is violated. PRODIST requirements must be observed, regarding grid voltages and power quality.

By following these steps in a structured and automated manner, the AIIA framework for the impact assessment of DG introduces a precise and interactive process. This approach ensures a wide range of analyses of how DG integration affects the distribution network under different conditions, contributing to informed rational decision-making and high-quality system performance. The OpenDSS code is a subroutine of the AIIA method.

4.5 Remarks on the Proposed Methodology

It is important to emphasize that, in applying the proposed method, the determination of the number of files for simulation via the OpenDSS software may vary, depending on the organizational strategy of the circuit elements. To illustrate this point, refer to Fig. 2, where low-voltage lines, denoted as “Lines-LV,” and medium-voltage lines, denoted as “Lines-MV,” are separated into distinct files for running in OpenDSS. Alternatively, one may choose to create a single file named “Lines” during the method implementation, incorporating both medium and low-voltage lines.

Another critical consideration is that utility databases, despite containing a wide range of information to be provided in a standardized manner, may exhibit certain inconsistencies. For instance, the absence of voltage data for a transformer or a typo in the connection location of an element may result in non-convergence or yield erroneous results in the power flow calculation performed by OpenDSS. These inconsistencies need to be analyzed and addressed individually in each case to enable power flow simulation.

5 Case Studies

To demonstrate and analyze the implementation of the proposed methodology in this study, a Python script was developed. This script is used to conduct two case studies aimed at identifying new investment areas in the distribution network and integrating DG. These studies are based on simulations of two Brazilian feeders currently in operation, modeled using BDGD:

  • Feeder SRS09 from CEMIG, the primary electric distribution company in the state of Minas Gerais;

  • Feeder RMN09X1 from Neoenergia COELBA, a distribution company in the state of Bahia.

After applying the proposed method and obtaining the required files for simulation via OpenDSS, each feeder undergoes a technical analysis to assess the level of DG integration in different regions.

It is important to highlight that the case studies presented in this section relate to the assessment of DG hosting capacity on the network to prospect new investment areas. However, the feeders generated via the proposed methodology can be used in other types of studies, considering different levels of DG connection and observing various electrical impacts.

5.1 Initial Analysis Definitions

The analysis of the impacts resulting from the integration of DG adheres to the limits outlined in the CEMIG Technical Standard (CEMIG, 2018) and the Neoenergia COELBA Technical Standard (Neoenergia, 2016). Voltage levels and branches loading across the network are evaluated for each scenario. Following the utilities’ methodology when issuing Opinion Access.Footnote 4 (or Access Information (Schmidt et al., 2021; Castro et al., 2023)), DG’s impacts are studied under the most adverse conditions from their perspective: zero load along the entire feeder and maximum generation at the DG units. In this scenario, where there is no consumption within the network, all power flows toward the initial feeder bar (the connection bar at the Electrical Substation entry point). As per the local utility standards, the power factor for photovoltaic generators must be considered as 1.00 (i.e., unity) at its peak value, which is the value used in these studies.

Regarding the voltage levels, they should fall within 95% and 105% of the nominal operating voltage of the system. As for the loadings on the segments caused by the current injected with the connection of the DG, considering the network at no load, they must not exceed 40% of the nominal cable carrying capacity. This 40% limit aims to replicate a typical approach already adopted by the Brazilian concessionaire CEMIG when requesting a DG connection. However, it is important to highlight that this is a dynamic criterion that can be adapted according to the management policy of the systems.

Additionally, to obtain a comprehensive study, as per the utilities’ interests, situations with network loads are also assessed. Hence, for each feeder under study, the results of three different scenarios are presented:

  • Scenario I No DG and presence of loads for the month of highest demand;

  • Scenario II With DG and no loads (the primary scenario of interest for the analysis of prospective areas for DG installation);

  • Scenario III With DG and loads for the month of highest demand.

Scenario I represents the baseline case of the study, denoted as pre-operational points (POP), simulating the distribution network in its original state from the BDGD, with all the existing loads connected. In the other two scenarios, alterations are made to the original system by connecting the DG units. Within the presented results, the highest demand value is gathered from the BDGD along with an hourly curve of the month. That is, although the presented results are not dynamic, the methodology is suited for time-series simulations.

Lastly, it is worth noting that for cases with DG presence (Scenarios II and III), the installations in two different regions of each feeder are also considered, illustrating an initial study of prospective areas for investment in the system.

5.2 Feeder SRS09

The first feeder analyzed is identified by the code SRS09 and is part of the Santa Rita do Sapucaí electrical substation, located in the municipality of Santa Rita do Sapucaí, in the state of Minas Gerais. This substation has a transformer with a capacity of 25 MVA and voltages of 138/13.8 kV, and it supplies power to five feeders. The average total loading of these feeders in 2021 was approximately 13.48 MW.

For feeder SRS09, the highest loading occurred in March. This feeder is the second largest in terms of length and load, serving non-urban areas in the city of Santa Rita do Sapucaí and also supplying power to the urban region in the municipality of São Sebastião da Bela Vista. After applying the methodology proposed in Sect. 4, the necessary files to simulate Feeder SRS09 using OpenDSS were obtained. Figure 4 compares the diagrams of the feeder from QGIS (Fig. 4a) and OpenDSS, using the files obtained through the proposed methodology (Fig. 4b). The electrical substation (ES) is also portrayed in these figures. As observed, the real feeder’s structure is faithfully replicated via OpenDSS.

Fig. 4
figure 4

Feeder SRS09: circuit diagrams via a QGIS; b OpenDSS

Full size image

In Fig. 5 the two regions of Feeder SRS09 selected for DG integration (Regions A and B) are depicted. This figure also highlights the types of conductors comprising the feeder, with their characteristics detailed in Table 1, where AAC stands for All Aluminum Conductor and ACSR stands for Aluminum Conductor Steel Reinforced. Region A is situated 4.6 km from the substation in a semi-urban section, while Region B is approximately 14.5 km away and encompasses the municipality of São Sebastião da Bela Vista, traversing non-urban areas before reaching the urban territory. For Region A, the conductors at the connection point have a capacity of 272 A, while for Region B, the conduction capacity is 204 A.

Fig. 5
figure 5

Feeder SRS09: regions for DG integration and conductor types

Full size image
Table 1 Feeder SRS09: conductor characteristics
Full size table

The following section comprises analyses of the three proposed study scenarios for Feeder SRS09. It is important to note that Scenario I applies to both regions, whereas Scenarios II and III produce distinct outcomes due to the varying locations of DG within the network.

Figure 6 illustrates the feeder voltage profile at the bars of Feeder SRS09 concerning their distance from the electrical substation for Scenario I. The red lines denote the prescribed limits, falling between 95% and 105%. The highest voltage within the network is 1.04 pu, which is close to the electrical substation and aligning with the accepted voltage range. Upon closer examination of Scenario I, it becomes evident that some low-voltage load buses operate at voltages as low as 0.948 pu, slightly below the established minimum. Given the voltage profile of the feeder, it is anticipated that the network will experience an elevation in these values following the connection of DG.

Fig. 6
figure 6

Feeder SRS09: voltage profile—scenario I

Full size image

In Scenarios II and III, which involve the installation of DG in the predefined Regions A and B, we determined the maximum DG power the feeder can support in each scenario. The goal is to ensure that the upper voltage limit, set at 105% of the nominal voltage, would not be exceeded anywhere in the system upon DG installation. The maximum power values and injected currents for each scenario in both regions are summarized in Table 2. This table reveals that Scenario II, where the network operates without any load, allows for the least integration of DG in both regions. This outcome validates the selection of Scenario II by distribution utilities for conducting studies and issuing Opinion Access, given its representation of the most critical operational scenario. Furthermore, as depicted in Table 2, Region A exhibits significantly higher potential for DG installation compared to Region B (approximately 4.17 times higher).

Table 2 Feeder SRS09: DG injected powers and currents
Full size table
Fig. 7
figure 7

Feeder SRS09: scenario II voltage profile: a region A; b region B

Full size image

The voltage profiles along the network observed with the integration of DG according to the maximum capacity identified in each region for Scenario II are presented in Fig. 7 (Region A in Fig. 7a and Region B in Fig. 7b). As expected, there is an increase in voltages in this scenario compared to Scenario I. Since the goal is to assess the DG hosting capacity until a limit is reached, the graphs do not show an extrapolation beyond the upper power limit.

When the feeder is analyzed with the integration of DG and the presence of load (Scenario III), Table 2 shows an increase in the capacity of Region A to accommodate generation without exceeding voltage limits. In this circumstance, the capacity of Region A is 6.1 times greater than that of Region B. The power flow in this scenario, with DG in Region A, can be contrasted with the power flow in Scenario I, where the load is the same, but there is no DG. These power flow distributions are depicted in Fig. 8 (Scenario I in Fig. 8a and Scenario III, with GD in Region A, in Fig. 8b), where the line thickness on the network is proportional to active power flow. From these results, it is noteworthy that while Scenario II serves the analysis of the most critical case regarding voltage impacts, it does not consider other aspects of the operating network and may, in fact, be more conservative than necessary.

Fig. 8
figure 8

Feeder SRS09: real power flow diagram: a scenario I; b scenario III, with DG in A

Full size image

The values of injected current and relative percentage loading \(\eta \) at the DG connection points for Regions A and B are presented in Table 3. These values are the same for all three phases and are provided for these points as they are the most affected by the presence of decentralized sources. It is observed in Scenario II, which is of interest when issuing Opinion Access, that there are no loading values (\(\eta \)) exceeding 40% of the nominal current limit. Similarly, in Scenario III, there was no exceeding of the loading limit. However, it is worth noting that the values defined for the DGs in the two regions are merely illustrative, as the distribution company CEMIG limits the power of DG connected at medium voltage to 2500 kW (CEMIG, 2018).

Table 3 Feeder SRS09: crrents on DG connection points
Full size table

The decision regarding which region to select in a practical study involves additional considerations, such as the installation costs in various feeder locations. In the case of Feeder SRS09, for instance, Region A may present lower DG implementation costs compared to Region B, primarily because it encompasses non-urban areas, despite having a lower hosting capacity. It’s important to note that the calculation of total investment costs is complex and involves dynamic variables such as expenses related to specialized labor, equipment, and transportation, among others. The assumption of lower costs for Region A is based on the premise that non-urban terrains in Brazil tend to be less expensive than urban areas within a municipality.

Finally, based on the results presented and analyzed in this subsection, it becomes evident that the simulated feeder, using files from the proposed methodology for simulation with the OpenDSS software, enables the assessment of the impact of DG integration in different regions.

5.3 Feeder RMN09X1

The second analyzed feeder is identified by the code RMN09X1 and is associated with the Remanso Substation operated by Neoenergia COELBA, situated in the city of Remanso, state of Bahia. The substation is equipped with two transformers of 69/34.5 kV, one with a capacity of 10 MVA and another with a capacity of 5 MVA. This substation serves as the point of origin for five feeders, with a total average load recorded at 11.37 MW for the year 2021. Notably, the highest recorded load for this feeder occurred in January (data for all months are accessible through the proposed methodology). This specific feeder ranks as the second largest and load-bearing within the substation. It primarily supplies electricity to non-urban areas within the city of Remanso and extends its reach to provide power to the city of Pilão Arcado. Importantly, it should be highlighted that Feeder RMN09X1 is longer than CEMIG’s Feeder SRS09.

After applying the proposed methodology, the files for simulating the feeder RMN09X1 using the OpenDSS software are obtained. Figure 9 displays the diagram of the feeder RMN09X1, showcasing the two selected regions for the DG integration study: Regions A and B. The diagram in this figure also indicates the types of conductors used in the feeder under examination, with their characteristics detailed in Table 4, where “CCS" stands for “Copper Clad Steel". Region A corresponds to a rural area, located approximately 7.5 km from the substation, while Region B is situated about 62.7 km away in the city of Pilão Arcado, traversing rural areas before reaching the urban region of the municipality.

Fig. 9
figure 9

Feeder RMN09X1: Regions for DG integration and conductor types

Full size image
Table 4 Feeder RMN09X1: cnductor characteristics
Full size table

As done for the first feder, three scenarios were conducted for Feeder RMN09X1. In Scenario I, as shown in Fig. 10, a voltage profile plot along the network considering loads but without the installation of DG, can be observed. When compared to the previously examined feeder, SRS09, more buses approach the upper limit of 1.05 pu. Additionally, numerous buses fall below the lower limit, with some reaching values as low as 0.89 pu. These low-voltage points are primarily connected in the low-voltage network, explaining why various distribution transformers have been adjusted to tap values near the upper voltage limit of the network. This situation arises due to the substantial dimensions of Feeder RMN09X1, leading to significant voltage drops along its circuit.

Notably, the highest concentration of loads in this network is located between Regions A and B, as illustrated in Fig. 9, necessitating the operation of transformation and regulation equipment with tap adjustments ranging from 1.04 to 1.05 pu to compensate for these voltage drops. With these adjustments in place, an analysis of the network without load representation and DGs already reveals buses exceeding the upper limit, as shown in Fig. 11.

Fig. 10
figure 10

Feeder RMN09X1: voltage profile—scenario I

Full size image
Table 5 Feeder RMN09X1: DG injected powers and currents
Full size table
Table 6 Feeder RMN09X1: currents on DG connection points
Full size table
Fig. 11
figure 11

Feeder RMN09X1: no DG and no load

Full size image
Fig. 12
figure 12

Feeder RMN09X1: region A—voltage profile a scenario II, b scenario III

Full size image

From the characteristics of the Feeder RMN09X1, evidenced up to this point, it is possible to observe that the study of prospecting areas for the installation of new DG units requires careful analysis, since the presence of these sources, depending on the connection point, can contribute both to improvement or deterioration of the network voltages.

For Scenario II, considering Regions A and B for connecting DG, the values obtained for the maximum powers supported by the feeder are shown in Table 5. These values were obtained considering that the insertion of DGs in the analyzed regions could not cause the violation of the upper voltage limit of 1.05 pu for any new bus in the network in relation to the case without DG. By way of illustration, the same was done for Scenario III, which considers the presence of loads, with the result also shown in Table 5. The loads at the connection points of the DGs for these two scenarios in the two regions are shown in Table 6.

It is evident that when examining Region B of feeder RMN09X1, Scenario II exhibits a reduced capacity for DG integration compared to Scenario III. Conversely, for Region A, the capacity remains consistent between both scenarios and is notably higher in comparison to Region B. The reason for these results is that in Region A, the DG is located very close to the substation. Moreover, there are no loads between the substation and Region A, resulting in similar conditions for Scenarios II and III in this regard. As this feeder already experiences several points of overvoltages, an additional extrapolation is the stop criterion (besides the loading of the branches). Due to the proximity and stiffness of the substation modeled in OpenDSS (the source represented as a Thevenin equivalent), the feeder can accommodate the same amounts of power (3 MW) until an additional voltage limit is reached.

The voltage profiles along the network with the connection of the DG in Region A can be observed in the graphs in Fig. 12 (Scenario II in Fig. 12a and Scenario III in Fig. 12b). Upon examining these graphs, it is evident that there is an increase in bus voltages for both scenarios, as expected. However, it is important to note that the connection of DGs does not rectify violations of the minimum voltage limit. Instead, it contributes to the overall improvement of the voltage profile. Further in-depth studies in this regard could even be conducted using the feeder obtained through the methodology proposed in this work.

5.4 Remarks on the Case Studies

The results indicate that the feeder models and the proposed evaluation strategy enable the assessment of impacts related to DG integration into the network from various angles, including the influence of load composition operating concurrently with generation. The analyses conducted, with the detailed feeders in focus, highlight, for instance, that while increasing the distance between DG installations and the substation may lead to a voltage profile increase along the network, connecting them near load centers can be advantageous. This underscores the necessity for comprehensive and consistent studies, following the proposed methodology, especially as DG presence in electrical distribution networks continues to grow.

6 Conclusions

The increase in DG installations in Brazil represents a positive impact on environmental issues, given that most DG units come from renewable sources. However, it is essential to consider the limitations of the existing distribution networks for the sake of sustainable growth in the sector. The substantial expansion in recent years may have led to the onset of network congestion, making increasingly precise studies necessary.

In this work, a methodology is developed for modeling feeders utilizing data from the Brazilian Utilities Geographic Database (BDGD in Portuguese). These feeders are then used to perform OpenDSS software power flow simulations, demonstrating that trustworthy feeder models can be obtained from public information, since the electric parameters provided are consistent. Specifically, the voltages and currents presented are in pace with what would be expected when integrating DG in the distribution networks. The case studies presented indicate that this methodology enables the automatic modeling of real feeders currently in operation, providing a robust foundation for conducting crucial studies related to the subject matter of this research.

To validate the proposed approach, the paper evaluates the impacts of DG within the context of the Brazilian regulatory framework. The novelty of the proposed methodology lies in its data-driven approach, which involves processing data from the Brazilian utilities’ geographic database to develop an up-to-date and structured simulation of distribution feeders, using only free and open-source core tools.

Regarding future work, two main points are considered: (i) the inclusion of resources for optimizing the location and sizing of DGs along the feeder, and (ii) the application of computational intelligence techniques to improve the effectiveness of the proposed methodology by assisting in processing the information available in the BDGD. The latter involves identifying data patterns inserted by utilities to address inconsistencies in the generation of feeders for simulation via OpenDSS, as discussed in comment 3. Using these techniques, it is possible, for example, to automatically fill in missing or incomplete data.