Active Microgrids with Dispersed Renewable Generation and Their Power Quality Performance Challenges ^†

Abstract

The research and analyses presented in this paper are an attempt to prove the concept that a flexible and efficient energy transformation requires a gradual digitalization of the energy system, starting from the inside out, i.e., from the low-voltage micro- and nano-grids, which mostly integrate low-power photovoltaic power plants and consumers with similar demand profiles. This approach is supported by the two main advantages of these grids: they are almost similar in structure and they are scalable, the two characteristics indicating a possible successful digitalization. For this to happen, we need to study the problems in these grids and be aware of the technological maturity of the energy facilities. This paper (the first of several to follow) examines the electricity performance problems caused by the stochastic nature of solar generation. A technique for monitoring and predictive load analysis is proposed, as well as technical measures for implementing decentralised control.

1. Introduction

Small-scale RESs with a capacity of up to 10 MW are often referred to as distributed generation (DG). This definition is related to the stochastic nature of the generation process and its geographical dispersion. Small-scale RESs are not connected to the main transmission grid due to the high cost of transformers and high-voltage switchgear. They need to be connected to the distribution grid. Such connection of small RESs is known as distributed generation (DG) or dispersed generation. It is also called embedded generation, because it is built into the distribution grid. In such distribution grids, energy can flow from point to point, which seems to be a non-traditional and non-standard way of energy flow. This bidirectional energy exchange creates additional challenges in terms of power quality performance, and the efficient operation and protection of the distribution network.

The integration of distributed small RES generation (DRES) is based on several technical, economic and environmental benefits. DRESs are typically small in capacity due to their lower energy density and their dependence on the geographical conditions of a region. To estimate the share of electricity or power that is supplied by DRESs, we use the concept of penetration. We distinguish between “average penetration” (1), which we use to estimate carbon savings, and “instantaneous penetration” (2), which we use for all other cases, e.g., system control. In general, the maximum instantaneous penetration is much greater than the average penetration [1,2,3].

$$Average penetration={\displaystyle \frac{Annual energy from RES powered generators, \mathrm{kWh}}{Total annual energy delivered to loads, \mathrm{kWh}}};$$

(1)

$$Instantaneous penetration={\displaystyle \frac{Power from RES powered generators, \mathrm{kW}}{\frac{Power from RES powered generators, \mathrm{kW}}{Total power delivered to loads, \mathrm{kW}}}}.$$

(2)

Concept of Active Distribution Microgrid

The active distribution network is also called the “embedded generation distribution network”. In the past, distribution networks have had a one-way transmission of electricity, i.e., they are passive as there is no added generation capacity. Today, distribution networks are becoming active through the addition of DG that induces bidirectional flows of electricity in the networks. For a well-functioning active distribution grid, it is important to have consumption management, integrated DG management and an interconnection between them for flexible and smart operation and control. Subsequent digitalization would be achieved by satisfying the following requirements: adaptive protection and control; wide-area active control; advanced sensors and measurements; network control instrumentation; real-time network simulation; knowledge of data collecting and analysing through AI methods; and new and technically adequate distribution network design [1,2].

Microgrids and nano-grids are active distribution networks and are a mixed system of renewable generation capacity and various low-voltage combined loads. To the advantage of microgrids, the following features distinguish them from conventional generation:

• The fundamental advantage of microgrids for consumers is that they satisfy electrical and/or thermal energy demands locally. This means that they can obtain uninterruptible power, reduced losses, improved local reliability and local voltage support.
• Although they have a much smaller capacity than the large generators in conventional power plants, they are adaptable and scalable, two important qualities for building digital twins.
• A major advantage of microgrids is that they can be treated as a controlled unit within a power system.

To achieve the stable and reliable operation of microgrids, their structural configuration requires the consideration of several technical, regulatory and environmental issues. For example, the establishment of standards and regulations for the operation of microgrids in accordance with electricity supply and distribution companies, the low energy content of the used fuels and the dependence of the energy generation on the climatic changes must be taken into account. The smart micro- and nano-grids have several important technical and economic benefits for the electricity industry: these include the reduction of environmental problems, the improvement of the utility and reliability of electricity, the economic use of energy flows and the ability to solve market problems. As shown in

Table 1. Useful functions of microgrids according to the connection mode.

Stand-Alone Mode	Connected to the Grid Mode
Active and reactive power control of micro-sources for stable voltage and frequency at the end of the load line.	Monitoring system diagnostics through the collection of information from micro-sources.
Adapting load break/stop strategies using storage-enabled demand management to maintain power and bus voltage balance.	Performing technical and safety condition assessment, economic generation planning, and active and reactive power management of micro sources, as well as consumption management functions using the collected information.
Switching the microgrid to grid-connected mode after power is restored from the main grid, without affecting the stability of either grid.	Ensuring synchronous operation with the main grid by maintaining power exchange at the connection points.

, these benefits are due to very useful features.

The voltage quality indicators of electrical power in low-voltage grids are frequency, deviations, rapid variations, asymmetry, non-sinusoidality, short-term drop, short-term interruptions and long-term interruptions. The values of these power quality indicators are regulated in [1].

The production and consumption of electricity in any grid cannot be fully balanced. Loads and currents vary, and voltage fluctuations occur. The voltage in the power grids changes depending on the load, the length of the transmission lines, the operating modes of the generators and other factors. Even a reduction in voltage by the minimum 3–5% from the nominal voltage results in an increase in the current consumed by the consumers, and this adversely affects the performance of all electrical equipment, causing energy losses increasing. Thus, in operating conditions for power transformers, there arises a need to adjust and/or maintain within certain limits according to the changes that occur due to an uneven or asymmetrical load distribution. The solution of this task faces considerable technical difficulties, is time consuming as a process and is further complicated by logistical and organizational problems. The voltage regulation of power transformers in power plants is defined as technical activities to limit the voltage deviation from the nominal values of the buses on the consumer side. Voltage regulation is carried out in distribution networks in order to ensure the economical and reliable operation of the power equipment. It is carried out depending on the primary voltage level in a given section of the grid or in the case of seasonal load variations. The maintenance of the voltage at the connection points within the technically permissible limits is made according to a preset law of variation. The voltage regulation laws are defined by the condition of ensuring the most economical joint operation of the reactive power sources, the electrical networks and the connected consumers. Influencing factors on the choice of output conditions are local conditions, type of network, configuration of consumers, etc. It is known that voltage regulation is related to the balance of reactive power in the network and on the buses to the consumers. Voltage reduction is generally observed in sections with a reactive power deficit [4,5].

In active microgrids with dispersed renewable generation, which we consider, power quality metrics are frequently and sometimes systematically violated. The negative consequences that follow are of a technical, technological and economic nature. In the case of domestic consumers, they appear as the inability to use electrical appliances and the risk of defects and accidents [2,6]. In the case of non-domestic consumers, the negative consequences are related to the disruption of the technological process, the impossibility of the proper operation of electrical equipment, etc. In the case of DG, the negative consequences are the automatic shutdown of the power plant at elevated voltage values and the loss of funds and preconditions for accidents. When the voltage exceeds the permissible limits, the negative consequences are most often associated with overheating and accelerated wear of the equipment. Electric motors, transformers and other devices are designed to operate at a rated voltage. Voltage overshoot leads to increased currents, the overheating of windings, shortened service life and increased power and energy losses. When the voltage is reduced below the permissible values, the negative consequences are related to a reduced efficiency of the electrical consumers. The electric motors lose power, their heating increases and their efficiency deteriorates. Electronic devices malfunction, lighting performance suffers and load currents increase (at low voltages consumers, especially motors, start to consume more current to compensate for the lack of power, leading to the overloading of networks).

In practice, both centralized and local voltage regulation exist. In centralized regulation, permissible voltage levels for groups of consumers are maintained simultaneously at a supply node. Depending on the load schedule, three types can be distinguished: voltage stabilization, two-stage regulation and counter-voltage regulation. Voltage stabilization is used for consumers with a nearly constant commodity profile, where it is necessary to maintain the voltage within a certain range over a 24 h period. Two-step regulation is applied to consumers that have distinct voltage levels. This maintains two levels suitable for this commodity profile (minimum load mode and maximum load mode). Counter-regulation is used for users with a multi-stage and variable commodity profile. Local regulation implies the maintenance of voltage levels immediately at the bus of a specific user. It is difficult to define a precise boundary between the two types of regulation, since local regulation at individual nodes of the power system appears simultaneously and centrally to other consumers at adjacent nodes. Local regulation is divided into group regulation (when implemented for a group of consumers) and individual regulation (when implemented for a single consumer with specific energy demands).

In fact, regulation in power transformers is limited to the stabilization of the secondary voltage $U_2$. This regulation can be performed by modifying/disconnecting parts of the secondary winding $W_2$ from the grid (step regulation) or by modifying the magnetic flux of the transformer that crosses $W_2$, so called under load regulation (technically difficult). The most common way is by modifying $W_1$ or $W_2$, which is in effect a change in the transformation factor. It is expedient to disconnect individual parts to be made on that winding whose voltage changes during operation. This ensures an almost constant magnetic flux during voltage regulation, since $\Phi \equiv {\displaystyle \frac{U}{W}}$, [6,7]. Usually, tripping deviations are made on a high-voltage winding, since it has a larger number of turns and more accurate regulation can be implemented. In addition, the current in the high-voltage winding (HVW) is smaller. The switching of the HVW can be performed in two ways: either by disconnecting the transformer from the primary and secondary grid or by the regulation of $U_2$ during operation (regulation under load). The first way is possible in cases of consumers that allow the interruption of the electrical supply. These are grid transformers with small and medium power, which have three stages. The basic grade is equal to the $U_{2N}$, and the other two are within ±5% $U_{2N}$, relative to the windings matching the middle position. Each step is switched for 1 to 3 s. This type of switching is also performed when the load changes seasonally. The switching mechanisms are of a simple design, but are manually operated, hence the transformer shutdown and the grounding of the windings are necessary. The second way of regulation is used in power ESs, when there is a need to redistribute active and reactive power. With a large range of regulation, the phase voltage amplitude is varied independently of the load in any of three ways: remote automatic, remote manual or local manual control [4,5,6,7].

Actually, without special regulating equipment, an acceptable voltage regime can only be ensured in conditions where the cumulative voltage losses in the power network are relatively small, for example, in lines of small length or with a small number of power transformers to convert voltage levels. For all power transformer loads, the standard IEC 60076-7, the loading guide for power transformers, defines three operating modes: normal cyclic loading, long-time emergency loading and short-time emergency loading. Another standard that is used is the IEEE Std. C57.91TM-2011-IEEE Guide for Loading Mineral-Oil-Immersed Transformers and Step-Voltage-Regulators. Four load modes are defined: normal life expectancy loading planned loading beyond nameplate rating, long-time emergency loading and short-time emergency loading. First, we define the operating mode for each specific case when secondary voltage regulation is required. Then, we follow the recommendations in the standards.

2. Technical Considerations

This paper analyses a common case of voltage performance distortion in a low-voltage network. Because of the operation of grid-connected PV plants with power 30 kWp and 56 kWp, systematic violations of the power quality indicators are observed for domestic consumers connected to the same bus. According to the recommendations of Regulation No. 16-116 on the technical operation of power equipment, we evaluate the variation in the operating parameters of power transformers. This paper presents the results of the measurement and analysis of the power quality indicators in a low-voltage distribution grid with two small-power photovoltaic plants connected to it. End users, PV plants and storage systems use the same grid. The objective of the study is to evaluate the impact of generation processes on power quality performance and on the proper allocation of loads to each phase.

Electricity supplied to end users and producers (PV power plants), respectively, needs to be of a certain quality to ensure the reliable and sustainable operation of the electrical equipment, supply lines and cables. The power quality indicators of electrical energy is regulated by [1], where they are strictly defined: continuity of supply, voltage deviation, voltage fluctuation, frequency, constant magnitude and sinusoidal shape of the supply voltage. In fact, these indicators cannot be strictly constant because they are influenced by a number of systematic and random factors and vary around certain values. Voltage deviation has unpleasant consequences for both consumers and electricity suppliers. For consumers, voltage deviation leads to the following negative impacts: increased processing time; reduced productivity; increased production costs; disrupted processing; increased raw material consumption and production rejects; and the shortened life of insulation, consumers and power system components.

For electricity distribution companies, as electricity providers, the voltage deviation leads to large financial losses due to the disruption of technological processes and the automatic disconnection of consumers. Another commonly violated indicator is asymmetry. Asymmetry in these grids can be short-term (occurring with asymmetrical short circuits) or long-term (occurring with asymmetrical voltage regimes of the power source, incomplete grid regimes caused by line and/or transformer phase tripping or uneven load distribution).

Since this type of grids feeds so-called identical consumers (with similar load patterns), voltage regulation in step-down substations in these cases is realized in one of the following ways: under-load voltage regulation or off-load voltage regulation. The transformers under consideration have three steps of regulation: the middle one is equal to and the other two to ±5% $U_{2N}$. The switching is performed on the HVW to distribute the currents symmetrically. Switching off parts of the windings causes an unequal distribution of the magnetic flux dissipation. A cross component appears in the magnetic induction vector, which interacts with the current flowing through the coil and creates electromagnetic forces that tend to move the coil in the axial direction. These forces are very dangerous when the transformer is in operation, placing it in a near-short-circuit condition, and can cause the secondary winding to be destroyed.

The switching and regulating processes are carried out with the transformer disconnected from the primary and secondary grid to prevent these phenomena. It is possible to regulate under load with a current-limiting reactor. This is the so-called fast-acting switching of a number of windings during the operation of the transformer, the Janssen principle, which is also carried out in a strictly defined sequence. The current-limiting reactor used is calculated for a short-time current load (~0.1s), with both its terminals connected to one stage of $W_2$. The resistance of the current-limiting reactor is practically zero, since its two halves are switched oppositely. Smooth regulation under load is accomplished by an under-magnetizing shunt through which a portion of the operating magnetic flux is shut off. One special feature that is often neglected, even when switching only one stage, is that a large inductive dissipation resistance occurs. For this reason, during switching, the current in $W_2$ increases further. This must necessarily be considered during operational switching because it leads to an increase in the electric arc flash time.

The compliance of electricity quality indicators for consumers is one side of the issue. The other relates to the power and energy losses during the often-demanding regulation processes. It is well known that the transformer efficiency is obtained at such a load when the constant and variable losses become equal. The load factor (3) and the efficiency (4) are determined by the following relations [6,7,8]:

$$k_{Load}=\sqrt{{\displaystyle \frac{P_{id}}{P_{sh.c}}}};$$

(3)

$$\eta =1-{\displaystyle \frac{P_{id}+k_{Load}^2.P_{sh.c}}{k_{Load}.S_N.{\cos \phi }_{Load}+P_{id}+k_{Load}^2.P_{sh.c}}},$$

(4)

where $P_{id}$ represents idle losses, $P_{sh.c}$ is short circuit losses, $k_{Load}$ is the load factor and $S_N$ is the rated/nominal total power.

To evaluate the energy efficiency and economy of a transformer during load variations, the so-called annual efficiency is often used. This concept is understood as the ratio of the energy supplied by the transformer to the secondary grid in one year to the energy consumed in the same period.

3. Results and Discussion

In the case under consideration, the period is one week, since load cycling is found for this type of grid. The measurements have been carried out in the area of a transformer station, with a low-voltage terminal, in a meter board supplying domestic consumers. The supply was made at a low voltage of 0.4/0.23 kV with aluminium non-insulated conductors that were type AC. The transformer in this grid was type TM-400 with the following catalogue data: $20/0.4 \mathrm{k}\mathrm{V},{∆P}_{id}=0.99 \mathrm{k}\mathrm{W};∆P_{sh.c}=5.92 \mathrm{k}\mathrm{W};I_{id}=1.8\%.I_N;u_{sh.c}\%=5\%$. The measurement was taken in real time with a portable power quality analyser according to [1] (according to IEC 61000-4-30: Power Quality Measurements), which allowed the detection and tracking of potential grid faults. The observed and recorded quantities in the three phases were the following: phase and line voltages, currents, power (active, reactive and total), supply voltage frequency, voltage deviation, harmonic components of voltage and current, total harmonic distortion of voltage and current, voltage and current asymmetry, rapid voltage changes, flicker and voltage dips and breaks. Additionally, all quantities for flicker, distortion factors and rapid voltage changes, as well as their voltage related events (DIPS), were observed. The measurement period was one week from 13 to 19 December 2024. In order to evaluate the measurement, the permissible values according to [1] were considered. These values provided a description of the electrical characteristics in the power supply grids. In a low-voltage microgrid under consideration, the nominal voltage is 230 V.

For the measurement period, the frequency of the supply voltage was almost constant, remaining in the range of 50.05–49.94 Hz. Significant deviations from the permissible values for phase voltages were observed. Under normal operating conditions, 95% of the 10 min average effective supply voltage values should be within ±10%U_N for any one-week period, i.e., in the range 207 to 253 V. The variation in the phase voltages for the three phases L1, L2 and L3 is presented in

Table 2. Phase voltage minimum, average and maximum values [V].

Phase Voltage, V	Value
	Minimum	Average	Maximum
L1, V	173.5	231.4	259.8
L2, V	183.5	232.4	257.1
L3, V	171.5	229.1	253.2

and Figure 1a–c. The following colour code was used in Figure 1 for better visualisation: cut-off values below (-10%) and above (+10%) the nominal voltage are represented by horizontal red lines; average values are in red; maximum values are in blue; and minimum values are in green colour. The minimum voltage values for the three phases were below the recommended limits. The maximum values for phases L1 and L2 exceeded the permissible values.

3.1. Flicker

Rapid voltage changes are rated by the flicker value and its severity over a long-time interval. According to [1], under normal operating conditions, rapid voltage changes should be no greater than 5%U_N. Changes of up to 10%U_N of small duration may occur several times per day under some conditions. The flicker should be less than 1 relative units (r.u) for 95% of a period of one week.

In Figure 2a–c, the variation in the flicker over the measurement period for the three phases L1, L2 and L3 is presented. A red horizontal line marks the permissible flicker level. For phase L1, the mean value for 95% of the records was 2.16, at a norm of 1 r.u. For phase L2, the mean value for 95% of the records was 3.78, at a norm of 1 r.u. No deviations above the permissible values for THD_U% and THD_I% were found in the measurements that was carried out.

3.2. Asymmetry

According to [1], the asymmetry should be within 0 to 2% for 95% of a one-week period. At a norm of 2%, the asymmetry had an average value of 3.34% for 95% of the measurements. To make the analysis more precise, we present the correlation between voltage level and long-time flicker for each phase. The relationship between the observed metrics for phase (L1) is presented in Figure 3a, for phase (L2) in Figure 3b and for phase (L3) in Figure 3c. The graph shows the time range over which the indicators worsened. In the hours from about 10 a.m. to 3 p.m., the voltage and the flicker in the grid increased. This time interval matched the maximum solar power output of the PV systems connected to it. The combined action of these negative phenomena led to power outages (the dips in the curves marked in green), shown in Figure 1 and Figure 3.

At times when there was no solar radiation, the grid voltage and flicker were within the norms. This is the problem of voltage regulation. It is necessary to work with a Janssen regulator, which takes almost 5–6 h to operate. During this time, consumers experience poor voltage quality and power losses, which increase the risk of accidents and failures. In the context of the need for digitalization, a methodology and the corresponding technical implementations for the selective management and order of operations are clearly necessary.

In this study an automated under-load adjustment system with remote operation capability was examined. Four important quantities were monitored. $I_{st}$: The operating rated current. This is the current that flows through the secondary winding of the power transformer $W_2$. $U_{st}$: The operating voltage between two nearby switched stages of $W_2$. $U_{st MAX}$: The maximum allowed voltage between steps. $P_{st N}$: The rated power of the switching step ($P_{st N}=I_{st}.U_{st}$). The limit value of $P_{st N}$ is the maximum power that ensures safe switching. IEC 60214 regulates the procedure for type testing and determining this switching power limit.

Implementing centralized and decentralized control requires load forecasting, actual solar generation resources, capacity assessments and electricity price data. Load forecasting and renewable energy generation depend on collecting and precisely processing large data sets. When combined with energy storage for sustainability and reliability, installation and operation costs increase. However, the benefits of forecasting should outweigh the additional costs. Currently, forecasting mainly focuses on PV power generation and prices. Load forecasting is rarely considered in microgrids [9,10].

In order to monitor the load variation, we chose a technique that is minimally affected by load variation, is relatively simple to calculate, allows for online monitoring, is more reliable in data collection and is hence convenient for the needs of predictive analysis. According to [11], the power factor, ($\cos \phi$), is not only an indicator of efficient operation, but can also be used as a marker of balance between active and reactive power. Any load variation or potential fault would result in a change in the grid parameters. Consequently, the power factor would also change. The nature of the load (inductive or capacitive) cannot be ascertained by $\cos \left(\phi \right)$, since $\cos \left(\phi \right)=\cos (-\phi )$. Therefore, $\sin \phi$ was used, i.e., $\sin \left(-\phi \right)=-\sin \left(\phi \right)$ and this made it possible to separately monitor potential faults and the effects of load variation. The diagnostic criterion $r\left(t\right)$ will only signal if there is a change in a previously defined normal operating condition with nominal values as follows:

$$r\left(t\right)={\displaystyle \frac{\sqrt{\frac{1}{\tau }{\int }_{t-(\tau /2)}^{t+(\tau /2)}{[u_1\left(t\right)-1/\tau {\int }_{t-(\tau /2)}^{t+(\tau /2)}{u}_1\left(t\right).dt]}^{2}dt}}{\sqrt{\frac{1}{\tau }{\int }_{t-(\tau /2)}^{t+(\tau /2)}{[i_1\left(t\right)-1/\tau {\int }_{t-(\tau /2)}^{t+(\tau /2)}{i}_1\left(t\right).dt]}^{2}dt}}}X\sqrt{{\displaystyle \frac{1}{\tau }}{\int }_{t-(\tau /2)}^{t+(\tau /2)}{[p\left(t\right)-1/\tau {\int }_{t-(\tau /2)}^{t+(\tau /2)}p\left(t\right).dt]}^{2}dt,}$$

(5)

where $\tau$—is the observation interval; p(t) is the generated power, $p\left(t\right)=\sum _{i=1}^n[I_i\left(t\right).U_i\left(t\right)]$; and $i_1\left(t\right)$ and $u_1\left(t\right)$ are the instantaneous values of phase current and voltage, respectively.

The calculation algorithm comes down to eight steps:

• We determine the observation interval $\tau ={\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$f$}\right.}$.
• “Moving” the interval τ without overlap and detecting the instantaneous current and voltage values passing through zero, we record them as $t_i$ and $t_u$.
• We calculate the interval $∆t=t_i-t_u$. When $∆t<0$ the load is inductive. When $∆t>0$ the load is capacitive.
• Using the formula $\phi ={\displaystyle \frac{∆t}{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$f$}\right.}}.2\pi =2\pi f.∆t$, we calculate the angle $\sphericalangle \phi$.
• Using the expression $\xi =\sin \left(\phi \right)$, we calculate the function $\sin \left(\phi \right)$. The variation in the angle $\sphericalangle \phi$ is calculated using the $\alpha$ criterion as follows:

  $$\alpha \left(t\right)={\displaystyle \frac{1}{3}}.[\xi .\left(t-\left({\displaystyle \raisebox{1ex}{$\sigma $}\!\left/ \!\raisebox{-1ex}{$2$}\right.}\right)\right)+\xi \left(t\right)+\xi \left(t+\left({\displaystyle \raisebox{1ex}{$\sigma $}\!\left/ \!\raisebox{-1ex}{$2$}\right.}\right)\right)]$$

  (6)

  where $\sigma ={\displaystyle \raisebox{1ex}{$2$}\!\left/ \!\raisebox{-1ex}{$f$}\right.}$.
• We calculate the deviation of the angle $\sphericalangle \phi$ by using the $\beta$ criterion as follows:

  $$\beta \left(t\right)=\xi \left(t\right)-\alpha \left(t\right)$$

  (7)
• We calculate the criterion $r\left(t\right)$.

According to variations in the criteria $\alpha \left(t\right)$, $\beta \left(t\right)$ and $r\left(t\right)$, the intelligent monitoring system reaches the following conclusions:

• The deviation of the criteria $r\left(t\right)$ is an indicator of the occurrence of a fault in the generation unit.
• If $\alpha \left(t\right)$ changes, but $\beta \left(t\right)$ and $r\left(t\right)$ do not change in the observed time interval, then it is the load variation.

Depending on the arrangement of RES generation, sensors and control devices, combinations of variations in the different parameters $\alpha \left(t\right)$, $\beta \left(t\right)$ and $r\left(t\right)$ can be given as a marker function for disturbed power quality performance and can control the operation of thyristors in the transformer control circuit. If one of the thyristors fails, it is possible to stop the control process, but this would be a first-order signal to the automated control system, and the central power supply would be shut down. The next step in automating the voltage regulation process is to provide control of the following quantities and parameters: regulation limits, overvoltage control, undervoltage control, zero-sequence voltage control and load current control.

4. Conclusions

Based on the measurements and analysis, it can be concluded that the power quality worsened during the monitored period. The most common problem was voltage deviation below the allowed range. Unacceptably low values ranging from $-20.2\%.U_N$ to $-25.4\%.U_N$ were recorded. The flicker was 1.16% to 2.78% above the acceptable range. The asymmetry was 1.34% above the recommended range. Additionally, no deviations were detected regarding voltage frequency or total harmonic distortion (THD). The most unpleasant and unfortunately common phenomenon occurs when solar energy production is at its peak, alongside flicker and voltage deviation. This automatically leads to the tripping of circuit breakers and the disconnection of power lines, along with all the negative consequences that follow. The problem is that this occurs daily during the period of the most intense solar generation, which occurs 156 to 260 days per year. Degraded voltage quality can lead to damage to appliances and electrical equipment, as well as accidents, breakdowns, lost profits and financial losses. Properly controlled microgrids increase electricity reliability by reducing power outages and their duration. Electric utilities should view active microgrids as controllable loads that can contribute to peak shaving during peak demand by reducing their own consumption and disconnecting noncritical loads. The authors are continuing their research in several areas, including the selection of appropriate short-term storage systems in active microgrids under known load profiles, as well as the predictive analysis of solar generation in these grids. The optimization objective is to minimize total losses in the power grid.