Impact on Voltage Regulation in Medium Voltage Distribution Networks Due to the Insertion of Photovoltaic Generators

Abstract

The objective of this paper was to analyze the impacts caused by the operation of voltage regulators in electrical distribution networks and to evidence the number of operations in the face of short-duration voltage variation caused by the high intermittency of the connected PV generators. A real LV and MV feeder was used, modeled in OpenDSS software, based on normative standards, adjustments, and technical maneuvers strategically used by the local utility. The analyses considered the temporal variations for the photovoltaic generators and different load demand profiles connected to the feeder. The feeder was submitted to the demand curves varying the load percentage, framing it in high and conventional (nominal) load according to the profiles of consumers and prosumers connected. The simulations made it possible to observe the exacerbated elevation in the number of maneuvers performed by the voltage regulators of the network. The single-phase voltage regulators stood out by the elevation of control operations, causing premature wear of the PV generation equipment connected to the most loaded phase. It was observed that discrepancies in the power flow in the lines and the voltage levels at the busbars. The creation of strategies and decisions to correct these impacts caused to transformers and regulators is possible.

1. Introduction

Each day the technology of renewable energy sources improves and favors the evolution of distributed generators, especially photovoltaic ones, facilitating their massive connection to the electrical distribution network. When connected to the grid, they can cause changes in the normal state of operation, causing variations in the voltage levels of the grid [1]. These variations are intensified by the high intermittency of the local solar radiation and due to the massive connection of the generators to the feeder [2,3,4,5,6].

The problems with the voltage levels and elevation of technical losses of the low voltage distribution grid can be intensified when the interaction between the grid and the small PV generators happens, as seen in [7,8,9,10]. This occurs when a massive connection to the grid of PV generators occurs in a distributed and/or concentrated manner in a single phase. It causes significant impacts on voltage levels, with the appearance of short-duration voltage variations, in addition to excessive voltage unbalance in the power grid.

It was observed in [11] that the massive connection of centralized or distributed PV generators along the low and medium-voltage electrical distribution networks could cause rapid power variations in the feeder, affecting the behavior of voltage regulators, capacitor banks, and the load tap changer—LTC. The equipment is important for grid operation and essential to perform the control and maneuvers to correct the feeder voltage levels during the system operation [12].

The voltage regulator equipment inserted in the distribution system faces difficulties in maintaining the normative voltage limits of the electrical grid when interacting with distributed photovoltaic generators. This presents itself as a “power quality” problem faced by the utilities. The relentless pursuit of consumer units for increased electricity consumption, coupled with more expensive energy tariffs, increases the demand for the installation of distributed photovoltaic generators on the electrical grid. This massive connection makes the grid more complex and, consequently, harder to maintain control of the normative voltage limits, making it difficult to operate the voltage regulator equipment inserted in the feeder [13] due to short duration voltage variation (SDVV), which is also a problem generated by the connection of PV generators to the electrical grids. According to PRODIST module 3 [14], these problems are expressed as the network responds to critical sags or rises in the feeder’s effective voltage value.

If the power utility uses technical operation artifices, such as adjusting the transformer ratio via the transformer’s TAP, it increases the voltage in the secondary to reverse the sags caused by the line impedance. This type of procedure offers benefits to the consumer units connected at the end of the feeder. However, it can result in the appearance of overvoltages on the feeder, especially when the loads connected to it are idle at the moment of the effective contribution of the connected PV-distributed generators. To correct the problems of deregulation of voltage levels, devices such as TAP changers are used, allowing a physical or automatic change in the transformation ratio of transformer equipment, regulating the voltage levels in the secondary service network to the limits established by the National Electric Energy Agency [15].

2. Finding Voltage Regulation Problems

2.1. Impacts on Voltage Regulation

The massive connection of PV generators to a distribution feeder can generate problems with the operation to normative standards, causing problems with quality and electric power supply to consumer units connected to this feeder [7,9,10,16]. Among these problems, we highlight the violations of the supply voltage levels of the loads and the increase in technical losses due to the high flow of power circulating in the feeder [17]. These problems hinder the correct operation of the equipment that composes the electrical grid, raising mainly the performance levels of voltage regulator devices, contributing significantly to their premature wear, reducing their useful life, and raising maintenance costs [7,8,9,10,17,18].

After analyzing Figure 1, it is seen that the voltage levels along the feeder naturally tend to suffer “voltage drops” as it moves away from the electrical substation. This occurs in distribution and transmission lines due to electrical losses caused by the capacitive and resistive effects along the feeder. These operational phenomena are considered and sized in electrical network projects with the objective of maintaining the voltage levels in the operating range of the normative standards. In addition to the regulating equipment implemented during the operation of the power grid, technical maneuvers are used, such as the technical adjustment of the transformer’s TAP, to raise the voltage along the feeder and correct the phase shift of voltage levels at the end of the lines. These operational strategies are important for the control and compensation of voltage levels due to voltage drops along the circuit [10].

However, this maneuver is performed by a team from the local utility, which travels to the region where the electrical substation presents problems in voltage levels and mechanically alters the transformer’s TAP, occasionally raising and/or lowering the voltage level in the secondary. This procedure can trigger other voltage problems when there is the presence of distributed photovoltaic generators connected to the feeder [19] since, in this scenario, not only the substation transformer would supply electrical energy to the loads connected to the feeder but also the PV generators. Thus, if a rupture in the voltage levels occurs beyond the normative standards determined in design by the utilities, it is again projected the need to perform “another” adjustment in the TAP of the voltage regulator of this section [19,20,21,22]. For the most part, voltage regulator equipment does not have automatic TAP adjustment in local Brazilian utilities due to the high costs of structural modification of the electrical grid. However, voltage regulation is a vital maneuver for today’s grid operation scenario, where it is extremely important to perform the accommodation of the huge amount of distributed photovoltaic generators connected daily to the MV and LV distribution grid.

In order to perform these operations, the power utilities adopt several specific operational procedures, aiming to minimize costs and solve the problems of unbalanced voltage levels [13]. These strategies are based on a scenario of intelligent electrical grids well suited to solve such problems. However, conventional electrical distribution grids do not have a communication infrastructure, which, in this case, is extremely important for the operational functioning of the grid. Thus, more localized solutions are adopted, as a resource, using energy storage systems and in the control and variation of the dispatched power [21,23,24,25].

The addition of distributed generators connected to the traditional electrical distribution network can cause different problems [26], and the creation of the generated bidirectional power flow significantly impacts the distribution system when related to voltage regulation.

2.2. Voltage Variation in Steady State

By increasing the number of connections from distributed PV generators to a distribution feeder, distortion of electrical parameters related to voltage variation in a steady state can occur. These parameters are linked to the intensification of the reverse power flow, reactive power, and the variation in the load demand profile of the feeder, interfering with the voltage profile along the feeder, which is a direct consequence of the injection of active power from the connected photovoltaic systems [26,27].

In the Brazilian system, normative requirements are imposed on system users and concessionaires for the connection and operation of the conventional electrical distribution network, which can be consulted in PRODIST—Operating Procedures for Distribution Module 8 [28] when related to the voltage range of operation. In

Table 1. Connection points at a rated voltage equal to or less than 1kV. Source: [29].

Service Voltage (TA)	Read Voltage Variation Range 380/220 (Volts)
Proper	(350 ≤ TL ≤ 399)/(202 ≤ TL ≤ 231)
Precarious	(331 ≤ TL < 350 or 399 < TL ≤ 403)/(191 ≤ TL <202 or 231 < TL ≤ 233)
Critique	(TL< 331 or TL > 403)/TL< 191 or TL > 233)

, standard voltage values are presented, which are applied for different levels of operation of the electrical network. In this case, the reference values for low-voltage operations are illustrated.

In

Table 2. Standardized nominal voltage in the medium voltage network. Source: [29].

Service Voltage (SV)	Variation of the Read Voltage (TL) in Relation to the Reference Voltage (TR)
Proper	0.93 TR ≤ TL ≤ 1.05 TR
Precarious	0.90 TR ≤ TL < 1.05 TR
Critique	TL < 0.90 TR or TL > 1.05 TR

, it is possible to observe nominal voltage values for the medium voltage distribution network, with operating values within the base of 13,800 V. The traditional work values are given in per unit (PU). They must meet [29] through adequate, precarious, and critical voltage ranges at the point of delivery with read and reference voltage. These procedures are assigned to operate within power quality standards for the electrical network.

It can be observed in Table 2 that there are voltage variation ranges that are accepted during the operation of the electrical system without causing inconvenience or damage to the installations. It is worth pointing out that in the Brazilian electricity industry, there is the regulatory agency ANEEL, which arbitrates operating limits for the power utilities. When these reference limits are exceeded, the distributors, depending on the level of disruption and the recorded time of the problem, are subject to financial penalties in accordance with the compensation system regulated by law.

The planning of the electrical system must be carried out dynamically because, currently, there is a growing increase in the number of connections to the grid from intermittent energy sources, such as solar and wind. This changes the entire power grid operation, which was not designed to accommodate this massive connection of distributed generators, creating problems at any time of the year distribution systems. The normative standards for grid operation used by the utilities assume an accommodation profile for distributed generators with deficient levels of interconnection, recognizing that the “niche” that the problems facing this interaction are increasingly present in everyday life.

The distributed photovoltaic generation, intrinsically intermittent by-passing clouds, can raise the voltage at the generator connection point beyond the acceptable limits in standard. This variability in generation produces many important and challenging interactions between the distribution system and the connected PV generators. For example, the difficulty in managing voltage levels results in increased wear and tear on regulating equipment and protection systems [9,30].

Thus, it is paramount to perform simulation studies to evaluate the behavior of voltage regulators involving the mutual operation between the electrical power system, its controls, and communications enhanced with distributed photovoltaic generators. Since problems in the voltage levels of the feeder caused by this interaction are already a reality and intensified when faced with a massive connection of PV generators, it directly interferes with voltage regulation problems, increasing the effective number of operations.

3. Modeling the Electric Grid

3.1. Photovoltaic System Model

Figure 2 shows the schematic diagram used to model the photovoltaic system in OpenDSS electric power distribution system simulation software. It shows the variables that define the model and the respective components involved in the modeling process.

At the point of interconnection with the grid, active power is defined ${\mathrm{P}}_{({\mathrm{t}}_{0)}}$ which is a function of the irradiation (reference), temperature, converter efficiency, grid voltage, and the power at the maximum power point (Pmp) of the PV panel output. The power at the maximum power point Pmp is corrected with temperature, just as the inverter efficiency curve must be parameterized as a function of the operating temperature. The value of $Irradiance$ is determined by the location of the photovoltaic array, and the maximum value of the analyzed period should be considered [32].

3.2. Transformer Model

The transformer model for the OpenDSS simulation system presents itself as a supplying element of electrical energy, called single-phase and/or multiphase equipment. To model the transformer in a simulation of the electrical distribution system, it is necessary to adopt at least its basic parameters, which are listed in

Table 3. Basic parameters for modeling the transformer element. Source: [32,33].

Parameter	Description
phases	Number of phases (default is 3)
windings	Number of windings (default is 2)
XLH	Series reactance per unit (pu)
% Loadloss	Percentage of total loss based on rated load
% noLoadloss	No-load loss percentage based on rated load
Wdg	Winding will receive the following parameters
bus	Name of the bar to which the element terminal is connected
conn	Winding connection (star or Delta)
kV	Terminal (winding) nominal line voltage in kV
kVA	Terminal-rated power in kVA
TAP	Voltage per unit (pu) of the TAP used

, in addition to representing them with the nomenclature adopted in the OpenDSS software. Furthermore, OpenDSS makes it possible to model several varieties of connections, for example, a transformation system with two or more coils; then, you can use the connection (Star-Delta) as a standard model or also allow the use of connection (Star-Star), allowing the use of one or more phases.

3.3. Voltage Regulator Model

The voltage regulator is an element whose characteristic function is to control and monitor the voltage in the winding of the transformer to which it is connected. In electrical power systems, it is possible to parameterize the regulators for three-phase voltage operation or three regulators for each phase when using single-phase regulators. This voltage control is possible when the transformer’s characteristics are parameterized through TAP adjustment [32].

The voltage regulator has the operational purpose of maintaining voltage control and compensating for voltage drops in the power grid transmission line. It is defined as a control element by the RegControl element described in

Table 4. Basic characteristics for modeling the RegControl element. Source: [31,32,34].

Parameter	Description
transformer	Controlled transformer name
winding	controlled winding
vreg	reference voltage
square	Potential transformer transformation ratio
band	Voltage value that defines the band around the reference value
Reversible	Indicates whether the regulator can be switched to make regulation in the reverse direction
Revethreshold	Reverse power in kW for regulator direction reversal
revband	Bandwidth for operation in reverse mode

, which presents the basic parameters of the element.

In [31,34,35], it is shown the OpenDSS operating parameters configure the voltage regulator in reverse mode. This configuration, widely used in line regulators, is enabled on the Reversible command, operates when the power flow is inverted and is more significant than a threshold defined by the reveThreshold command.

Control of the voltage regulator can be accomplished by adjusting the preset parameters through the fixed TAP, as seen in Figure 3. In addition, it is necessary to set the voltage levels, nominal transformer ratio, and voltage range width. In OpenDSS software, by default, the amount of TAP of the transformer to be controlled is equal to 32, that is, 16 TAPs above and 16 below the neutral position.

The voltage regulator can be dimensioned to operate in the electrical distribution network from four basic configuration modes [36]. The first mode is presented in Figure 4, which shows the connection model that performs control of the voltage level in the secondary winding of the transformer to which it is connected. In this case, the voltage regulator is parameterized to perform operations as an LTC or line regulator when the monitored transformer is an autotransformer. It is worth pointing out that the winding to be monitored should not necessarily be the same as the one where the TAPs are. The second mode has a similar operation, and the difference is in the use of the PT (Potential Transformer) that can be connected to any bar, allowing the simulation with smart grid devices [32,37].

The third mode performs Line Drop Compensation (LDC) voltage level control, with line voltage drop compensation. This mode is typically used for monitoring voltage levels at a load or load center. The line resistance (R) and reactance (X) parameters must be supplied in volts, as well as the nominal current of the current transformer (CT) primary, which is used to calculate the line voltage drop. The objective is to ensure that the TAP of the regulating transformer undergoes regulation changes when taking the voltage drop reading along the line. The analog circuit connection model is shown in Figure 5, with the voltage drop, ΔVcommp, expressed to emulate the voltage drop in the line impedance, ΔVlinha, reflected the regulator circuit. The line voltage drop is expressed as the sum between the line resistance and the line reactance multiplied by the line current. In contrast, the voltage drop in the compensator analog circuit is expressed as the compensation impedance multiplied by the compensation current. To ΔVcomp emulate the line voltage drop in the secondary of the TP, ΔVcomp × ptratio must equal ΔVlinha. That is, it equals ΔVcommp to ΔVlinha by reducing the value of ptratio, which is the TP transformation ratio that converts the voltage in the controlled winding to the regulator’s control voltage level.

The voltage regulator can also operate in reverse mode, applied to line regulators, and is enabled in OpenDSS from the reversible command. Its operation is activated when the power flow is inverted and exceeds the limits defined by the revThreshold parameter. In this mode, you can define parameters similar to those defined in the previous modes, but with the rev prefix, for example, revband, revreg, revR, and revX. There are two important parameters to specify in this mode. One is revDelay, which corresponds to the time needed for the control initiation action in reverse mode to start once power flow in the reverse direction above the threshold is detected. The other parameter corresponds to revNeutral. When set to “Yes,” whenever the regulator enters reverse mode, the TAP position is fixed at the neutral position.

3.4. Feeder Description

A real feeder from the local distribution system was used, which supplies power through a primary electrical network in medium voltage (MV) and the secondary circuits in low voltage (LV), as shown in Figure 6. The primary network originates in an electrical substation (ES) with a transformer with nominal power of 5MVA, a transformation ratio of 69.0/13.8 kV, and supplies power on a “permanent basis” to five secondary distribution transformers. Among them, one belongs to a private consumer unit with 75kVA nominal power, and four with 112.5 kVA nominal power belong to the concessionaire; All transformers have a 13.8 kV/380 V transformation ratio and a “Delta-star” type connection.

The LV radial electric networks start from distribution transformers with an operating voltage level of 380 V phase-phase (three-phase line voltage) and 220 V phase-neutral single-phase voltage and present in their structure three-phase overhead lines that operate in unbalanced conditions, transmitting electric power in the 4-wire model to the connected consumer units.

For this purpose, a total of 107 buses (busbars and/or nodes) are allocated in their actual positions, arranged in order to perform the branches and the support of the electrical distribution network. The rules were adopted for the average distance between the utility bars, which is 35 m, while the distance for the connection branches, “utility-bar” and “consumer-bar,” varies between 5 and 20 m. The value of 20 m applies when the consumer unit is on the opposite side of the unilateral bus, characterized as a bilateral bus.

The number of consumer units connected to a single pole/busbar is six loads, according to the actual distribution. To make these connections, some factors were considered: such as the layout of the cables, the distance from the consumer units, and obstacles (trees, houses, roofs, among others). When the network cable layout was unilateral on the street, it was considered the connection of three consumer units on the same side of the pole and three consumer units on the opposite side of the pole. For the bilateral arrangement on the same street, the busbar connection, only three consumer units per pole (busbar) were considered. The electrical network presented in Figure 6 has in its totality 601 consumer units connected to different types of consumption profiles. From this total, 548 are loads with a residential profile and 53 with a commercial profile.

In

Table 5. Number of consumer units connected per phase in each transformer.

Identification of Transformers	Single-Phase Loads			Biphasic Loads	Three Phase Loads	Total
	Phase A	Phase B	Phase C	AB Phase	ABC Phase	Consumers
TRAFO-A	122	-	5	-	1	128
TRAFO-B	144	3	2	1	15	165
TRAFO-C	120	6	6	-	10	142
TRAFO-D	100	20	20	5	20	165
TRAFO-E	-	-	-	-	-	1
Total	486	29	33	6	46	601

, the electrical characteristics of the loads connected to the grid are specified, such as the type of connection, the number of consumer units per phase of their respective transformer, and the nomenclature of each transformer used. In addition to this, the total number of consumer units per phase is observed. It is important to emphasize that these are real data and the modeling in the software followed the same criteria established in the collected data.

The loads connected to the secondary network are constant power (PQ) and present a power factor equal to 0.92. The reactive power production was considered in minimum conditions for the secondary network, knowing that, in reality, this factor is disregarded by the utilities for the secondary distribution system.

The electrical parameters of the voltage regulators connected to the grid are presented in

Table 6. Voltage regulators.

Regulators	Number of Phases	Ref. Secondary Voltage	Band	Delay	R Line	X Line	Primary CT	Monitored Phase
REGTRA_AB	3	120	1	2	0.4	0.6	20	ABC
REGTRA_C	3	120	2	2	0.4	0.6	20	C
REGTRA_D	3	120	2	2	0.4	0.6	20	A
REGFASE_A	1	120	2	2	0.4	0.6	20	A
REGFASE_B	1	120	2	2	0.4	0.6	20	B
REGFASE_C	1	120	2	2	0.4	0.6	20	C

. All the regulators are three-phase configured as voltage regulators with Line Drop Compensation (LDC) and have a VT with a transformation ratio of 20:1, according to the regulatory standards of the utility [28]. The regulators allocated in the lines were configured in reverse mode to verify the behavior of the voltage levels downstream and upstream of the bus to which it is connected.

4. Methodology

For the simulations, the software OpenDSS (Open Distribution System Simulator) was used, developed by Electric Power Research Institute (EPRI), to carry out the analysis of the electric power system in a steady state in a Permanent sinusoidal regime—RMS.

The program performs most of the analyses in a Permanent sinusoidal regime (RMS), which is widely used to carry out the planning studies of electric energy distribution systems [32,34]. OpenDSS performs studies related to the electrical system, for load flow, a connection of distributed generators, and in the study of smart grid solutions and practices (Smart Grid).

The real MV and LV distribution network was modeled in the OpenDSS software through the normative technical parameters, characteristics of the real demand curves of the feeder in question, the loads (consumer units) allocated and configured for each connected transformer, in addition to the connection of photovoltaic systems. Figure 7 represents the modeled curve for the connected photovoltaic systems, obeying a solar irradiation profile with high variability and, for the loads, a load curve (demand) with average load was inserted, of the actual feeder in question, with the objective of configuring the software with technical and operational data and thus conducting the analyses as close to normality as possible.

The prosumers and traditional consumer units were allocated to their buses, and their load consumptions were assigned considering the regulations of the local utility in case of installation of distributed photovoltaic—PV microgenerators and minigenerators. Thus, it was possible to analyze the behavior of voltage levels through responses from monitors implemented at strategic points of the lines, transformers, and voltage regulators installed when large medium and low-scale voltage PV systems are connected to the feeder.

4.1. Description of Scenarios

4.1.1. Scenario 1 (Base)—Distribution System without PV Distributed Generation

In this scenario, a simulation of the electrical distribution network was performed in the absence of power from the photovoltaic generators. Energy analyzers (Monitors) were modeled and allocated in strategic positions to record the behavior of the power flow in the lines, voltage unbalance levels, and variation of the voltage regulator TAP position levels, in addition to other technical parameters that support the system analysis. It is worth noting that the demand curves were submitted to the feeder, varying the carrying percentage and framing it in conventional (nominal) loading and high loading according to the profile of the consumers connected to the feeder.

4.1.2. Scenario 2—Insertion of Single-Phase and Three-Phase PV Systems with Power Demands Varying with the Daily Profile

In this scenario, the PV generators were introduced at the system busbars, separating them between three-phase and single-phase cases. For both cases, the simulations performed were divided into the variation of load demand and photovoltaic power with daily profile curves. The separation between the cases of three-phase and single-phase connections aimed to separate the effects of both since unbalanced connections tend to increase voltage unbalance rates. Subsequently, both situations will be considered simultaneously. The active power supplied by the photovoltaic systems massively connected to the feeder was parameterized at 20% (single-phase), 35% (three-phase), and 55% (three-phase and single-phase) with the feeder demand curve at average load.

In

Table 7. Description of the distributed generators connected to the feeder.

PVSystem	Phases	BusBAR	Voltage (V)	Power (kW)	PVSystem	Phases	BusBAR	Voltage (V)	Power (kW)
PV1TRA	1	01BT_TRA.1	220	8
PV2TRA	1	02BT_TRA.1	220	6	PV20TRA	1	20BT_TRA.1	220	10
PV3TRA	1	03BT_TRA.1	220	7	PV21TRA	3	17BT_TRA	380	25
PV4TRA	1	04BT_TRA.1	220	10	PV1TRB	3	01BT_TRB	380	40
PV5TRA	1	05BT_TRA.1	220	9	PV2TRB	3	03BT_TRB	380	35
PV6TRA	1	06BT_TRA.1	220	7	PV3TRB	3	06BT_TRB	380	50
PV7TRA	1	07BT_TRA.1	220	10	PV1TRD	3	Busbar_CLIENTEBT_TRD	380	120
PV8TRA	1	08BT_TRA.1	220	8	PV2TRD	3	Busbar_CLIENTEBT_TRD	380	120
PV9TRA	1	09BT_TRA.1	220	10	PV1TRC	3	21BT_TRC	380	28
PV10TRA	1	10BT_TRA.1	220	11	PV2TRC	3	19BT_TRC	380	18
PV11TRA	1	11BT_TRA.1	220	2	PV3TRC	1	18BT_TRC	220	12
PV12TRA	1	12BT_TRA.1	220	6	PV4TRC	1	14BT_TRC	220	10
PV13TRA	1	13BT_TRA.1	220	10	PV5TRC	1	15BT_TRC	220	12
PV14TRA	1	14BT_TRA.1	220	8	PV6TRC	1	16BT_TRC	220	12
PV15TRA	1	15BT_TRA.1	220	9	PV7TRC	1	17BT_TRC	220	10
PV16TRA	1	16BT_TRA.1	220	10	PV1TRE	3	CLIENTEBT_TRE	380	50
PV17TRA	1	17BT_TRA.1	220	7	PV2TRE	3	CLIENTEBT_TRE	380	35
PV18TRA	1	18BT_TRA.1	220	6	PV3TRE	3	CLIENTEBT_TRE	380	35
PV19TRA	1	19BT_TRA.1	220	9	PV4TRE	3	CLIENTEBT_TRE	380	55

, it is possible to identify the PV generators that were introduced to the feeders and other characteristics such as phase, connected transformer, operating voltage level, and installed power. It is worth noting that the power of the generators followed a variation profile according to the percentages mentioned in the previous paragraph, making it possible to extract the maximum from the feeders to which these PV generators were connected.

5. Results

When analyzing the results, it was possible to observe electrical parameters close to the real operating characteristics, in voltage levels, active and reactive power, as can be seen in Figure 8a–c. As it is an unbalanced circuit, as seen in Table 6, where most of the loads connected to low voltage radiating circuits are concentrated in phases A and B, it was expected that the simulated results projected this high imbalance between the phases. Figure 8b,c highlight the values of active and reactive demand “requested” by the consumer units to the electrical substation, respectively. Figure 8c shows the values of reactive power contained in the system since, at low voltage, the consumers are connected to the feeder through a power factor of 0.92, justifying the unbalance contained in the circuit and the prevalence of reactive power in Phase A of the electrical network.

Figure 9a–c graphically present the active powers of the feeder observed by a monitor allocated at the output of the electrical substation when single-phase, three-phase, and two-phase distributed PV generators were connected to the circuit simultaneously. The installation of PV generators in consumer units followed an asymmetrical distribution profile, focusing on each prosumer’s power availability and compatibility. Even with the real unbalance of the feeder, it was possible to connect the PV generators without altering the technical parameters of the network, i.e., the voltage limits and the power quality profile established by the normative characteristics. Thus, the limits were not exceeded during the network operation, demonstrating that the voltage regulators were sufficient to maintain the voltage within the adequate energy supply levels.

In Figure 9a–c, it was observed that during the contribution period of the PV-distributed generators connected to the feeder between 5 am and 5 pm, there was power compensation from the substation compared to the base case. An unbalance between phases was observed, caused by the massive connection of PV generators to Phase A of the feeder since it has many consumer units connected. In Phase C, however, relatively few changes were observed throughout the day, remaining practically unchanged from the base case seen in Figure 8b.

Figure 9b, with only the installation of three-phase PV generators, the expected naturally occurred; all phases received balanced, active power, which caused an imbalance between phase C in relation to phases A and B, an unexpected situation when it occurs the connection of three-phase systems to the electrical grid. In this particular case, this discrepancy occurred due to the load distribution characteristics along the feeder; as seen in Table 5, phase C has the fewest connected consumer units, making it idler compared to the other phases. Justifying the reverse power flow in the line that occurs between 7 am to 2 pm and at varying peaks at 3 pm and 4 pm, portraying that during the operation of PV generators, there was excess active power in this phase.

In Figure 9c, it can be observed through the results of the behavior of the power flow in the feeder when single-phase and three-phase PV generators are connected to it simultaneously. In the period of the contribution of the generators, between 5:00 am to 5:30 pm, it is seen that there was excess power flow in the three phases of the circuit, highlighted in Figure 9c, with the presence of reverse power flow in phases A (−50 kW), B (−40 kW) and C (−15 kW) approximately. Furthermore, during the grid operation, unbalance between phases was observed between the period of the PV generator’s contributions, and the amplitude of the reverse power flow between phases decreased. System response signals the power contribution from the PV generators in this period.

The significant presence of reverse power flow reflects the excess power in the feeder’s distribution lines, linked to the demand profile with average load, influencing the grid’s electric power quality parameters. The unbalance in the voltage levels, which directly affected the functioning of the voltage regulators installed in the feeder, had to increase during the operation, and the adjustments in the transformer TAP to keep these variations in the voltage profile regularized. Coupled with this unbalance, the presence of short-duration voltage variations was concluded by the numerous attempts of regulators to regularize the voltage levels. When PV generators are included in a distribution network, the taps of voltage regulating devices automatically adjust to accommodate the voltage increase caused by these generators. However, suppose the PV systems are disconnected from the grid. The new tap positions may cause voltage drops, impairing the quality of power supplied to the loads since the voltage regulator actuation is slow, taking tens of seconds to recover the voltage to the desired value.

In Figure 10a–d, the behavior of the voltage regulators can be observed through the transformer TAP’s maneuvers for the base case without the connection of PV generators, only with single-phase generators connected, only three-phase and both simultaneously, respectively. For the base case, the TAP adjustment maneuvers that occurred during the daily period were justified by the natural voltage imbalance present in the feeder and remained at levels with more stable levels without characteristic changes. The greatest requests for adjustments occurred in the period of greatest load demand request to the feeder, observed at night.

When comparing Figure 10b to the base case, it is seen that the insertion of only single-phase PV generators caused high impacts when related to changes in transformer tap positions, since the tap adjustment maneuvers during the solar contribution period reached levels with high levels of regulation. When connecting three-phase PV distributed generators, Figure 10c, there were changes in TAPs, however, with less demand for adjustment requests, varying their positions between +3 and +4. Peaks were observed at 7 am, 10 am, and 3 pm, and regression to their positions was verified in the base case, demonstrating that there was a short-term voltage imbalance.

Figure 10d, on the other hand, shows the results when there was a massive connection of single and three-phase PV distributed generators simultaneously, highlighting the high rate of requests to voltage regulators to perform the adjustment maneuvers in the transformer’s TAP in order to regularize the changes in the grid voltage levels. This problem was caused by the excess of circulating power coming from the injection of power by the connected PV generators, contributing to the voltage unbalance in the lines and consequently increasing the number of operations of the voltage regulators.

However, it was possible to conclude that the impact generated by the photovoltaic systems connected to voltage regulation was small. This is due to the high three-phase short-circuit power of the electrical substation, in accordance with the fact that the automatic voltage regulator portrayed in the results of this article is located in the electrical substation identified in Table 6 as REGFASE_A, REGFASE_B, and REGAFASE_C. Thus, more intense PV generation variations would be necessary to change the tap position of the automatic voltage regulator until inferring a significant change in the patterns of threshold voltage levels.

Therefore, with the objective of portraying the high changes of operations in the positions of the TAPs of the other voltage regulator equipment of this feeder, it was necessary to carry out the simulations not only considering the insertion of distributed PV generators but also for different variations in the demand curve of the feeder. In addition to the scenarios assigned in Section 4.1, load demand curves were used, varying between low and medium loading, projecting a consumption of 20% and 40%, respectively, for each system evaluation period. In this way, the possibilities of simulations were increased, meeting the objectives of the study, seeking to verify the “stress” in the regulation of the voltage level in the distribution network caused by the massive insertion of PV generators distributed along the feeder.

Table 8. Number of voltage regulator operations.

Regulators	Base Case			Single-Phase GDFV			Three-Phase GDFV			Single-Phase and Three-Phase GDFV
PHASES	A	B	C	A	B	C	A	B	C	A	B	C
REGTRA_AB	7	3	2	6	4	3	8	5	4	9	4	2
REGTRA_C	7	3	2	8	3	2	12	8	3	12	4	3
REGTRA_D	6	6	6	10	5	1	6	5	2	10	2	4
REGFASE_A	5	3	3	6	3	4	8	6	4	8	3	2
REGFASE_B	6	4	2	7	2	2	4	4	6	9	6	6
REGFASE_C	6	4	2	7	2	2	4	4	6	9	6	6
TOTAL	37	23	17	44	19	14	42	32	25	57	25	23

depicts the number of operations performed by each voltage regulator connected to the distribution network in question. Through it, it is possible to verify the different behavior of the voltage regulators in view of the responses of the electrical system through the operations carried out to accommodate the profiles of PV generators connected to the network. In addition to the behavior, it is seen the importance of regulatory equipment for an electrical distribution network.

Based on the results of Figure 10, it was concluded that the largest number of operations carried out for voltage regulation by the regulating equipment occurred more effectively in the daytime, with the largest insertion of distributed PV generators with great oscillation in the generation. By merging the established scenarios with the variation of the load curves of each prosumer connected to the powered one, it was possible to obtain the results of the voltage regulators contained in Table 8.

However, it is worth mentioning that during the simulations, operations performed by some voltage regulator equipment were observed that are not related to the increase in the penetration of distributed PV generators to the grid—verified when the load demand curve of the feeder was 20% more than its installed power, being equivalent to the base case in question, during some periods of the day. In turn, the long-term voltage variation cannot be observed through the responses of the voltage regulators in question because it was attenuated by the injection of active power from the connected photovoltaic distributed generators and presents similar numbers of regulation operations.

However, as there was a progressive increase in the connection of distributed generators and the change in the behavior of the demand curve (average), the short-term voltage variation became quite significant. This fact occurred due to the great intermittence contained in the local solar irradiation curve, which provided the superposition of long-term voltage variations, directly contributing to the increase in the number of adjustment operations by the regulatory equipment. It is observed in Table 8 that mainly there was an increasing rate of installations of distributed generators with different single-phase and three-phase power levels simultaneously to PHASE A.

6. Conclusions

The injection of active power by photovoltaic generators causes an increase in nodal voltages in all phases if it is performed in a balanced way between the phases of the electrical system. These findings can be justified by the facts portrayed in this article, firstly by the special characteristic of this feeder under study, as it is an unbalanced network, with phase A being significantly more loaded than the others. For this reason, voltage regulators connected to Phase A are the most affected by voltage variations on the feeder.

For the case of connecting single-phase generators, it was possible to conclude that the voltages of the three phases of the system are affected, and not only the voltage of the phase in which they are installed. This happens due to the displacement of the neutral position, caused by the variation in the level of unbalance and the existence of mutual coupling between the phases. An aspect of great relevance is that the increase in the distributed generation in one of the phases can provoke the reduction in the voltage in another phase.

Distributed generators influence the actuation of automatic voltage regulators, causing additional device tap changes. The relevant positive point is that the action of the voltage regulator avoided the violation of the steady-state voltage limits as the active power injected by the generators increased. Therefore, in general, voltage regulators can allow an increase in the penetration level of distributed generation without violating the steady-state voltage limits.

The power increase in the photovoltaic generation reduced the voltage unbalance index when the three-phase generators and maximum load were considered. For single-phase generators, the voltage unbalance index increased significantly, reaching values very close to the normative limits. However, when multiple generators distributed on different buses and phases were considered, the increase in the level of unbalance was more modest when compared to the value calculated when the generators were concentrated in a single-phase bus. This shows, once again, the benefit of providing a balance between the phases when connecting the photovoltaic generators. In the period when the PV system provides maximum active power, there is an average load demand on the distribution system. In this way, reverse power flow situations are possible when the degree of penetration of photovoltaic generators is high.