Supply of MV Island with High-Penetration of Prosumer Renewable Energy Sources

Abstract

The rapid development of prosumer renewable energy sources (RESs) observed in Poland in recent years causes problems in distribution networks such as current amplitude and voltage asymmetry increases, power and energy loss increases, and reverse power flows, and related are voltage control problems, deterioration of energy quality, etc. Moreover, in the case of planned repair/maintenance works in the network and the need to supply energy consumers in an islanded MV grid, the problem of the correct operation of such a subsystem appears. This occurs when the power production by the prosumers’ energy sources at a given moment exceed the power consumption. In such a case, reverse power flows occur in MV/LV transformers, i.e., from the LV network to the MV network. This causes reverse power flow to the diesel generator, leading to its shutdown and, in extreme cases, to damage. The solution to this problem is to use a mobile system equipped with energy storage in addition to a diesel generator and an LV/MV transformer. An additional problem in the case of using a mobile system (diesel generator) to power an MV island is the islanded MV network grounding. Grid islanding changes the earth fault current and electric shock voltages. In general, MV networks in Poland operate as compensated, i.e., grounding transformers are used, the star point of which is grounded by a compensation choke. Unfortunately, in the case of powering an MV island from a mobile system, there is no real possibility of grounding the star point of the LV/MV transformer used there. This article proposes an algorithm of a diesel generator with an energy storage selection, including electric shock protection requirements verification, for the use in suppling energy via an islanded MV network.

1. Introduction

Operational work carried out by a distribution system operator (DSO) in an MV network usually involves the need to disconnect a fragment of the MV network from the rest of grid. MV subsystem islanding is possible by using disconnectors and reclosers (remotely controlled disconnectors), which are located in selected locations only in the MV network. This makes it impossible to disconnect, for example, one repaired pole or single line span. This means that typically a certain fragment of the MV network is separated, in which there are several to a dozen MV/LV substations supplying LV networks. In such LV networks, there may be up to several hundred LV energy consumers. This means that an MV subsystem islanding may require power supply for up to several thousand LV recipients. Currently, this is performed by independent power supply of individual MV/LV substations using diesel generators operated by a power team (operator employees). Taking into account the formal procedures accompanying the power grid work, this is a complex logistical challenge, requiring significant human resources (power teams and dispatchers) and material resources (diesel generators).

A solution to the above problem may be use of a single diesel generator supplying a separate part of the MV network via an LV/MV transformer. Due to the specific nature of MV network operation, this requires an analytical evaluation of the possibility of safe operation of such an MV island each time. This results from the fact that MV networks in Poland operate as so-called compensated, in which compensation chokes connected to grounding transformers located in 110 kV/MV substations are used.

MV subsystem islanding leads to a reduction in the value of the earth fault current and a reduction in the risk of electric shock at the point of the fault. On the other hand, supplying a part of the MV network from a diesel generator involves the need to use an LV/MV transformer, which cannot be grounded at its star point. This means that for each MV island supplied from a generator, the protection against electric shock must be analytically assessed.

An additional problem that has emerged in recent years results from the dynamic development of prosumers’ RESs located in the LV network. A very large number of these sources means that during periods of high solar radiation the power generated by these sources exceeds the load, which leads to power flows from the LV to MV network and to the diesel generator. If the generator is equipped with reverse power protection, the generator is switched off, with a consequence being the of loss of power supply for electricity consumers. If the generator is not equipped with such a protection system it may be damaged, with the same effect for consumers.

This problem can be partially solved by carrying out work in months when the solar radiation is low, i.e., during winter in Northern Europe. But in practice, this significantly limits the possibility of carrying out maintenance work in the MV distribution network. A more advantageous solution is a hybrid analytical–technical system consisting of the following:

• In the technical area, such a system should consist of a diesel generator, an electrical energy storage, and an LV/MV transformer. This system is referred to in this publication as a hybrid mobile station (HMS), as shown in Figure 1. This system should be able to balance active power for at least several hours, i.e., charge and consume it depending on the power generated by prosumers’ RESs.

• In the analytical area, the system should provide the network operator with information about the possibility of safe operation of a separated MV and LV network.

In order to conduct such analysis, information is required on the structure of the separated MV network, ground fault protection settings in the 110 kV/MV substation, residual current resulting from grid compensation and historical information on the load profiles of the energy consumers.

Data on consumers’ load profiles are obtained from remote reading meters (individual smart meters (ISMs)), which have been gradually installed in Poland over the past few years. This results from implementation of Directives (EU) 2019/944 [1] into the Polish Energy Law, according to which 80% of consumers should have this type of meter by 2028, and all consumers by 2031. Regardless of ISMs, operators also install remote reading meters (station smart meters (SSMs)) in MV/LV substations.

If the islanded subsystem consists on the LV network only, historical measurements from IMSs should be used. And if the islanded subsystem consists of MV and LV grids (contains MV/LV substations), then historical measurements obtained from SSMs should be used. This type of case is considered in this publication.

In general, hybrid power systems (based on various technologies) are known and have been used for years. They are flexible, what makes them applicable to a wide range of applications. They are utilized in systems operating as permanent or temporal electric islands. They are used temporarily because of failures or maintenance work. In the first group are marine systems (vessels), while in the both groups are industrial, telecommunication, transportation, electric power, etc., systems.

Hybrid systems provide different operation configurations, thus proper control and management makes fulfilling the most efficient demand or control goals possible. Energy storage and diesel generator sizing is discussed in [2,3]. In a typical configuration, diesel generators operate as prime movers, being the primary source of electrical energy, while energy storage shaves the load peaks. Additionally, such a hybrid system with energy storage can limit the amount of diesel fuel used.

In the case of marine applications there are many different systems combining both mechanical and electrical drives with a hybrid power supply. Such solutions are based on the cooperation between mechanical, combustion generating sets and energy storage, which is intended for electrical supply [4,5,6].

In the case of land applications the discussion related to a diesel generator with RESs in an islanded grid focuses mainly on: the energy source’s control (often fewer or more RESs are considered) [7,8,9,10], the optimal placement of RESs in the grid, the optimal parameters of RESs in the grid [11], the islanded grid static and dynamic features, and the RESs placement and operation costs [12,13,14,15]. Also, results of tests conducted in a real power system are presented.

Remote areas and islands isolated from the main electric power system ensure reliability and stability of power supply using diesel generation. Due to the high purchase and transportation costs of diesel fuel, isolated communities are exploring other options for efficient and reliable power supply. Renewable energy sources can reduce operating costs [12]. When, because of various reasons, it is impossible to supply loads from the hybrid system, the use of local renewable sources and a small battery control is proposed [16,17]. This way of grid control requires the system observability and controllability, including ability to control prosumer (local) RESs.

In real systems, tests are conducted to verify the ability of such subsystem islanding and resynchronization. However, stand-alone operation including voltage and frequency control is also discussed [18,19]. It is indicated that the islanded grid’s low inertia and inverter features significantly impact protection, control, and grid stability, especially during large transients.

Also discussed are islanding partition and scheduling strategies which the system operator can conduct before grid maintenance to optimize islandic operation [20].

The problem under consideration is similar to microgrids operating islanded. A comprehensive review of almost 200 publications on microgrids and hybrid systems in [21] shows that island operation of microgrids is not typical and requires special attention in balancing active power to maintain the desired frequency value, which is why many authors do not address the topic of island operation of microgrids. Publications dealing with island operation of microgrids, on the other hand, assume the need for a complex central or hierarchical energy management systems which controls all of the energy sources in the grid. This paper assumes that the active power balance and frequency value will be controlled by only one source, which is the HMS. The literature focuses usually on a well-defined system, i.e., SG, PV, and energy storage, with known data and parameters. The presented paper considers the HMS sizing for a real grid, i.e., with measurements containing gaps of data and errors.

2. Materials and Methods

Evaluating the possibility of supplying power to a separate part of the MV network, the MV subsystem load profile should be determined. The assumed period of operation of the HMS and its maximum power should be considered. Determining the load profile also allows us to determine the amount of energy delivered by the diesel generator during its assumed operating time. This allows for us to also determine the necessary storage capacity. With some approximation, this can be performed using historical measurements carried out in the distribution network. In general, loads can be considered as the MV/LV distribution transformers and as the loads connected directly to the MV grid. Simultaneously, energy sources, e.g., photovoltaic farms or wind farms, connected to the MV network should be included in the considerations. Then, using the historical power/energy profiles of these devices, the HMS production/consumption profile can be determined.

In the case of the DSO under consideration, most MV/LV substations are equipped with an SSM. Measurements in the stations are performed on the LV side. The DSO has access to these measurements within the AMI (advanced metering infrastructure) system. In general, hourly values are measured in MV/LV distribution stations, including the value of active energy transmitted to the LV network (A+) and to the MV network (A−). Therefore, the “+” or “−” sign indicates the direction of current flow through the MV/LV station. In addition to the energy values, instantaneous and average values of currents and voltages are also recorded. These values can be used to determine the HMS load profile. But, the key factor here is the availability of measurements for individual MV/LV substations.

Energy measurements are sufficient to determine the capacity of the electrical energy storage in the HMS. For this case, the active and reactive energies measured in all MV/LV substations located in the MV island should be summed as follows:

A_HMS = Σ_t Σ_i{k_(i)·(A+_(i,t) − A−_(i,t))],

(1)

where

• i  —  i-th MV/LV substation charged by HMS with LV/MV;
• t  —  t-th measurement hour;
• k_(i)  —  factor for i-th MV/LV substation, resulting from current transformer ratio, to which a given SSM is connected.

In turn, to determine the HMS load profile, current and voltage measurements, instantaneous or average, can be used. It should be noted here that the use of instantaneous or average values carries a certain uncertainty. If average values are used, there is no information about the maximum value that occurred in an hour. The situation is similar with the instantaneous value, which is recorded at the time of measurement, i.e., every 1 h. To limit the uncertainty resulting from the measurements, the current value that is the greater of these two values can be used. Therefore, in order to determine the HMS load profile the powers of the individual phases for each MV/LV substation should be summed up as follows:

S_HMS(t) = Σ_i[k_(i)·(V_L1(i,t)·I_L1(i,t) + V_L2(i,t)·I_L2(i,t) + V_L3(i,t)·I_L3(i,t))],

(2)

where

• i            —  i-th MV/LV substation charged by HMS with LV/MV;
• t            —  t-th measurement hour;
• V_L1, V_L2, V_L3  —  instantaneous or average voltage in phase L1, L2, and L3;
• I_L1, I_L2, I_L3    —  instantaneous or average current in phase L1, L2, and L3.

It should also be noted that the determined apparent power S is on the low-voltage side, so it will be slightly higher on the medium-voltage side because it will take into account the losses incurred in the MV/LV transformer. Additionally, the HMS load profile determined in this way does not include losses incurred in the MV island transmission lines. However, neglecting these losses introduces an error in the safe direction. Since the current measurements available in the AMI system do not contain information on the flow direction (current module measurement available), the above method of determining the load profile is correct only for power flows towards the LV grid. In order to determine the flows towards the MV network, active energy measurements A+ and A− can be used. If the values both are non-zero then during the measurement hour there was a power flow in both directions. In such a case, the direction of current flow cannot be unambiguously determined. Then, it is assumed that the direction of flow is determined by the higher value. In turn, the unambiguous direction of current flow will only be for cases in which one of the values is zero. That is, if A+ > 0 and A− = 0 then there is assumed flow to the LV grid. And, if A+ = 0 and A− < 0 then there is assumed flow to the MV grid.

In general, it is not easy to precisely determine the load profile of the HMS supplying an MV island with high RES saturation. On the one hand, it is helpful to use historical measurements, which are described above. However, the currently used measurements do not provide weather information that would be helpful in assessing the share of RES generation in the power balance. And since most RESs located in the LV network are photovoltaic installations, information on cloud cover and insolation is crucial. Therefore, the HMS load profile determined on the basis of historical measurements should be verified after taking into account historical cloud cover and forecasted cloud cover for the day of the planned work.

An example of determining the HMS profile based on historical data for the MV island is presented below. The separated fragment contains 11 MV/LV distribution substations (Figure 2) for which historical measurement data are available. The analyzed fragment of the network consists only of MV overhead lines, the poles of which are marked in Figure 2 with dots.

3. Determination of the Load of the Mobile Power Station

As mentioned above, historical measurement data are needed to determine the HMS load profile. The period of measurement data analyzed covers the annual period from 1 October 2022 to 30 September 2023.

• It should be noted that, in general, the measurement data are not error-free or there are gaps in the measurements. For the considered network fragment some irregularities were noticed that required explanation. For example:
• In the case of distributed substations 21,030, 22,040, and 22,782, there was a multiplier of 200, which in reality turned out to be equal to 80, because 400/5 A/A current transformers are installed at these stations. Clarifying these inconsistencies required checking the DSO’s paper documentation.
• In the case of distributed substations 21,030, 22,153, and 22,285, there were flows towards the MV grid during night hours. It should be noted here that each of the 11 distribution substations contain only prosumer photovoltaic installations. This was considered as a measurement error because there is no power flow into the MV grid during nighttime hours.
• There are gaps in the measurement data that concern single hours, whole days, or some number of whole days. This paper presents a simplified approach in which the gaps are filled by evenly divided energy values, calculated as the difference in first (after the gap) and end energy values measured. In the considered case the missing measurements are rare (<1%).

Despite the inaccuracies in the measurement data described above, the HMS load was assessed. As previously mentioned, the direction of power flow for the HMS was determined by comparing the active energy consumed (A+) and emitted (A−) by each of the 11 MV/LV distribution substations located in the considered area. Below, Figure 3 and Figure 4 show load profiles for two sample MV/LV distribution substations. Out of the 11 distribution substations located in the considered area, only 1 (Figure 4) is characterized by only power transmitted towards the LV network. In the remaining cases, energy flows occur towards the MV grid, as for substation 21,116 (Figure 3).

As a result of summing up the load of all 11 MV/LV substations, a profile of the HMS load that would occur in the measurement year was obtained (Figure 5). Figure 5 shows the apparent power value S for each hour of the day for each day of the year. Analyzing the obtained profile, it can be seen that for most of the year the power balance has a negative value, which means that it is not possible to operate the grid with a traditional diesel generator on these days. Power flows towards the LV grid (with a few exceptions) are observed from the second half of November 2022 to the beginning of February 2023. This means that powering the analyzed part of the MV network using a classic solution (only a diesel generator) would only be possible during this period.

The data presented are historical. Therefore, while planning maintenance/repair work in an islanded grid, the weather forecast for the planned day of work should be included.

Limiting planned work to specific days of the year is a major obstacle for DSOs. In turn, the use of an HMS equipped with electricity storage significantly increases the freedom in planning these works. Nevertheless, DSOs need information on the maximum power that an HMS should have and the expected energy balance during the planned work. This is important in the context of selecting the storage capacity and its permissible charge level.

Considering the analyzed fragment of the MV network, it is possible to determine the maximum power that an HMS should have and what energy it should supply or consume on a given day. To determine these values, it was assumed that the power supply to the MV island lasts no longer than 6 h. Therefore, the HMS operating time was assumed to be from 9 am to 2 pm. For such a limited time during the day, the maximum power supplied by an HMS should not exceed 652 kW (Figure 6), and the maximum power consumed should not exceed −148 kW. Taking into account the previously adopted assumption that the rated power of the HMS is 630 kVA, the day with a load of 652 kW should be excluded from the planned work. However, such a load occurs only once a year. In other cases the maximum power achieved is below 600 kW.

The energy required to be supplied or consumed by the HMS during 6 h of operation is greater for the supplied energy and exceeds 3.96 MWh (Figure 7). In general, this is easy to achieve because it can be provided by the diesel generator, which is present in the HMS system. In turn, the energy consumed by the HMS for the most demanding day reaches 2.12 MWh. Therefore, the DSO who wants to carry out work on any day during the measurement year should use an HMS with an energy storage with capacity equal to 2.12 MWh. On the other hand, to carry out planned work on the MV grid it needs a mobile solution that will be able to relatively quickly connect, for example, to an MV overhead line pole. In such a case, a solution built on a truck comes into play. This means that the location of the diesel generator and energy storage on such a truck will limit the storage capacity. In turn, in order not to further limit the energy storage capacity, the LV/MV transformer could be built on a trailer.

Analyzing the measurement data presented above, it can also be seen that using the classic solution, i.e., only with a diesel generator, the DSO can carry out work on only 155 out of 365 (42.5%) days in the measurement year (Figure 8). On the remaining days in the assumed working hours (from 9 am to 2 pm) there is at least one hour in which excess power is generated. As a result, this would result in reverse power on the diesel generator and its disconnection. Therefore, the DSO needs a solution that will ensure the balancing of active power in the MV island, including the possibility of energy collection.

4. Effect of MV Island Separation on Electric Shock Protection

The second issue related to separation of a fragment of the MV network is the risk of shock hazard assessment, i.e., the calculation of the touch voltages that may occur during an earth fault.

As mentioned in Section 1, separating a fragment of the MV network changes the method of network earthing. Prior to separation, the MV network operates as compensated, using a compensation choke installed at the 110 kV/MV substation. The MV network is often islanded without the compensation choke. Additionally, the LV/MV transformer of the HMS has no real possibility of earthing the grid star point. So, this creates a network operating with an isolated star point, which changes a ground short-circuit current I”_k1.

Such a short-circuit current in an islanded grid has to be evaluated (calculated). To do this for overhead transmission lines with conducting supports (poles) without a disconnector the standard EN 50341-1 [22] can be applied. This standard defines the permissible touch voltage V_D (

Table 1. Largest permissible contact voltage V_Tp (V_D) in MV power lines according to EN 50341-1.

tF [s]	0.05	0.1	0.2	0.5	1	2	5	10
VTp (VD) [V]	735	633	528	204	107	90	81	80

) as a function of the earth fault duration (de facto as setting time of the earth fault protection). In the considered full MV grid (not islanded), the earth fault protection is set to ∆t = 1.2 s, while the single-phase short-circuit current (residual current) is equal to I”_k1 = 17.5 A.

The electric shock protection verification for MV power lines should be performed by calculation of the earth voltage V_E (Figure 9) for each MV uninsulated pole using its earth resistance R_s and comparing the voltage to the permissible touch voltage V_D, which is shown as follows:

V_E ≤ 2 V_D,

(3)

V_E = I_E·R_s,

(4)

For this purpose, the earth electrode resistances of the poles in the MV island are needed. Unfortunately, often only resistances for poles with a disconnector or recloser are available (

Table 2. Earth electrode resistances of poles with disconnectors located in the analyzed MV island. Values obtained from measurements.

Disconnector No	Rs	VE	2 VD for t = 1.2 s	Fulfilled Shock Protection
	[Ω]	[V]	[V]
942/II	7.44	2.3	200	Yes
198/II	28.2	8.7		Yes
975/II	1.56	0.5		Yes
290/II	1.2	0.4		Yes
1642/II	5.28	1.6		Yes
1274/II	33.48	10.4		Yes
420329	3.96	1.2		Yes
422115	12.96	4.0		Yes
964/II	20.28	6.3		Yes
1490/II	1.2	0.4		Yes
263/II	9.12	2.8		Yes
421785	69.72	21.6		Yes
961/II	3.24	1.0		Yes

).

Using the resistance R_s and the earth fault current I_E, the earth fault voltage V_E can be calculated. By comparing it to the permissible touch voltage V_D, the electric shock protection ability in the islanded grid can be verified.

Notice that some of the resistances are high. The highest one reaches 69.72 Ω. This means that for residual current equal to I”_k1 = 17.5 A a ground voltage reaches a value equal to 69.72 × 17.5 = 1220 V. It leads to the conclusion that electric shock protection requirements are not fulfilled. Such a conclusion can be based only on information about the pole earthing resistance R_s. In practice, some additional electric shock protection measures can be used in the pole earthing which can lead to the fulfillment of electric shock protection requirements.

In the considered MV grid the declared time of operation of the earth fault protection is equal to 1.2 s. For this time the permissible touch voltage is equal to V_D = 100 V (Table 1). This value is significantly higher than the calculated earth voltages V_E (see Table 2). The voltages are calculated for earth current in the MV islanded grid equal to I_E = 0.31 A. The current calculation is discussed below. But, in the considered case, i.e., for each pole of an overhead line with a disconnector or recloser located in the MV island, electric shock protection is met if 1.2 s setting time of earth protection is used in the HMS.

It should be noted that the above verification of the effectiveness of electric shock protection was carried out only for poles with a disconnector. For the remaining uninsulated poles in the MV island there is no information on earthing resistances. For this reason, as well as in the situation of low reliability of measurement data, where there is no information on the application of additional protection measures it is impossible to determine the effectiveness of electric shock protection on the basis of comparing the earth voltage with the permissible voltage.

Another of the important issues related to the neutral point grounding of the MV network is the arc extinction limit current. For a 15 kV network operating with an isolated neutral point, the VDE 0228, teil 2 standard (1987) defines this value as equal to I_limit = 35 A. The short-circuit current should not be greater than the arc extinction limit current, which is shown as follows:

I″_k1 ≤ I_limit.

(5)

Therefore, in the considered case, while defining the MV island the calculated earth current must not exceed the value of 35 A.

Another issue to consider is the impact of the presence of cable lines on the MV island. The cable lines significantly impact the values of earth currents due to a much higher capacitance in comparison to the overhead lines. A characteristic property of MV cable lines is their construction. They contain shielding conductors. The MV networks operate with shielding conductors earthed on both sides. In such systems during short-circuit, part of the short-circuit current flows through the shielding conductor, reducing the value of the current flowing through the earth at the point of short-circuit. This reduces the earth voltage. In practice, when determining the contribution of the shielding conductors to the flow of short-circuit currents a reduction factor r is used. The factor defines what part of the short-circuit current flows through the pole earth and is shown as follows:

I_E ≤ r·I″_k1,

(6)

The reduction factor takes values below 1 and varies for different shielding conductor cross-sections. For example, for MV cables with polyethylene insulation, the reduction factor is equal to 0.25 for a Cu 50 mm² shielding conductor, 0.4 for a Cu 25 mm² shielding conductor, or 0.55 for a Cu 16 mm² shielding conductor.

Given the above, it is necessary to determine the earth current for the MV island. This current can be determined using the symmetrical components method. It results from the following series connection of impedances: zero-sequence Z₀, positive-sequence Z₁ and negative-sequence Z₂. The single-phase short-circuit current is calculated as equal to the following:

I″_k1 = I_E = (c·V_n·√3)/|Z₀ + Z₁ + Z₂|,

(7)

where

• V_n  —  nominal voltage of network;
• c   —  voltage factor (1.1 for MV) [23].

In networks operating with an inefficiently grounded neutral point, the equivalent impedances for the positive-sequence Z₁ and the negative-sequence Z₂ are many times smaller than the impedance for the zero-sequence Z₀. Therefore, the positive- and negative-sequence can be neglected, which leads to the following equation:

I”_k1 = I_E = (c·V_n·√3)/|Z₀|,

(8)

The equivalent impedance for the zero-sequence of Z₀ depends on how the neutral point of the network is earthed. For a network operating with isolated neutral points, it is equal to the following:

Z₀ = 1/jωC₀ = 1/jB₀,

(9)

where

• C₀  —  transmission line zero-sequence capacitance,
• B₀  —  transmission line zero-sequence susceptance.

It should be noted that the value of power line capacitance (susceptance) used in the calculation of earth fault current is not the same as the capacitance (susceptance) used in the calculation of power line charging current. For short-circuit calculations, the zero-sequence of capacitance C₀ or susceptance B₀ is used, whilst for power line charging current calculations the positive-sequence of capacitance C₁ or susceptance B₁ has to be used.

The power lines used in MV networks can be overhead or cable. In the case of the MV island under consideration, there are only overhead lines, whose zero-sequence capacitance C₀ depends on the mutual arrangement of wires and their height above the earth. A comparison of the capacitance values B₀ and B₁ for an example of three arrangements of wire spacing on an overhead line support (Figure 10) is shown in

Table 3. Equivalent scheme parameters of selected MV overhead lines. Type of conductor: AFL6-70.

Conductor Support	R0	R1	X0	X1	B0	B1	I″k1
[-]	[Ω/km]	[Ω/km]	[Ω/km]	[Ω/km]	[μS/km]	[μS/km]	[A/km]
Minimum height of hanging wires above the earth (5.5 m)
Triangular	0.58	0.43	1.55	0.38	1.54	3.01	0.044
Flat 1	0.58	0.43	1.52	0.39	1.63	2.93	0.047
Flat 2	0.58	0.43	1.52	0.39	1.47	2.91	0.042
Maximum height of hanging wires above the earth (16.25 m)
Triangular	0.58	0.43	1.55	0.38	1.22	3.01	0.035
Flat 1	0.58	0.43	1.53	0.39	1.26	2.92	0.036
Flat 2	0.58	0.43	1.52	0.39	1.27	2.91	0.036

. Since the value of B₀ also depends on the height of the hanging wire, values are shown for the minimum (5.5 m) and maximum (16.25 m) hanging wire used by the DSO.

By analyzing the obtained susceptances it can be concluded that the zero-sequence susceptance is 40% to 60% higher than the positive-sequence susceptance. In addition, changing the height of the hanging cable only affects the zero-sequence susceptance. This change reaches up to 29%, depending on the arrangement of the wires on the pole.

In order to analyze the impact of the arrangement of the wires on the support (pole), the hanging of the wires above earth, and the sag of the wires, the MV island was modeled in the PowerFatory 2022 software from DIgSILENT (Gomaringen, Germany). The software used gives the possibility to take into account the geometry of the arrangement of the wires, converting it to electrical parameters of sections of the transmission line. This is particularly important from the point of view of relatively accurate determination of the capacitance of individual power line sections, which depends significantly on the distance of the wires from the earth.

In general, the DSO does not have information in its IT databases about either the arrangement of the wires on the poles or the height of the wires hanging above the earth. In the case considered, this information was obtained from flyovers of the analyzed part of the network MV by reading this information from graphic files of individual spans (Figure 11). The data provided show that all MV line sections occurring in the analyzed area have a triangular arrangement of wires, and the total length of all line sections is 8.007 km.

In networks operated with an isolated neutral point it can be assumed that the earth fault current will be the same at each point of the analyzed network. This is due to the dominant value of the zero-sequence, which leads to a reasonable simplification involving the neglection of the positive- and negative-sequence in the calculation of earth fault current. Therefore, the evaluation of the effectiveness of electric shock protection can be carried out independently of the network structure, i.e., there is no need to map the network topology, as information on individual spans and cable line sections is sufficient.

The occurrence of the same earth current at any point of the MV island under analysis is proved by calculations carried out in the mathematical model. Figure 12 shows four different locations of short-circuit, and in each of them an earth current equal to 0.31 A is obtained. This value is relatively small but is confirmed by the kilometric parameters.

The results obtained are also not affected by the location of the HMS connection. This is confirmed by an analysis of the various HMS connection locations shown in Figure 13a. Regardless of the connection location, the earth fault current has the same value (Figure 13b–e). So, from the point of view of electric shock protection, the HMS can be connected at any point on the MV island.

Short-circuit current is clearly influenced by the height of the hanging wires and the sag. In the general case, the lower the hanging wires, the higher the line capacitance (Table 3) and consequently the higher the earth current.

The effect of sag on the values of earth currents is relatively small. Using the considered network model, it can be found that the difference in earth current for the sag from the power line flyover performed at 19 °C and sag calculated at 40 °C differs by 0.6% (Figure 14). On the other hand, assuming no sag (sag equal to 0 m), the difference to the case with maximum sag does not exceed 3%. Thus, in the calculation of electric shock protection, in order to simplify the calculation, it is acceptable to assume the maximum overhang of wires.

A greater influence on the values of earth currents is the hanging height of the wires. Taking into account three different heights, 5.5 m, 10 m, and 16.25 m, differences in earth currents reach up to 29%. The calculations have been performed for an exemplary AFL-6 overhead line of 35 mm² with a length of 1 km (Figure 15).

At the same time, it should be noted that calculated values of the earth short-circuit currents were not verified with measurements in the real grid. This is because the DSO has not implement the proposed method yet.

Evaluation of electric shock protection can be performed based on the earth fault protection setting for the section from which the MV island is supplied. In this case, this setting is 17.5 A, with an operating time of 1.2 s. Thus, if the MV island’s earth current is less than this setting, it can be assumed that electric shock protection is satisfied. In the MV island analyzed, this condition is met because the earth fault current is equal to 0.31 A (then voltages V_D (Table 2) are very small), which is far less than the residual current (17.5 A), confirming that the electric shock protection is met.

On the other hand, if there is a case in which in the MV island the earth current is greater than the residual current, the operation time of the HMS earth fault protection should be calculated from the following relationship:

(V_{D(t, island MV)})/(I”_{k1(island MV)}) ≥ (V_{D(t, station110 kV/MV})/(I_{(res. station 110kV/MV)}),

(10)

where

• V_{D(t, island MV)}      —  is the voltage resulting from the characteristics for the permissible touch voltage V_{D(t, island MV))} = f(t_{prot Island MV}) (Table 1) for the time of operation of the earth fault protection installed in the HMS, for which the checked inequality is satisfied;
• V_{D(t, station 110kV/MV)}  —  is the voltage resulting from the characteristics for the permissible touch voltage V_{D(t, station 110kV/MV)} = f(t_{prot. station MV}) (Table 1) for the operating time of the earth fault protection installed in the section where the MV island is located;
• I_{res station 110kV/MV}    —   is the minimum residual current for the section from which the MV island is supplied during normal operation;
• I”_{k1(Island MV)}       —   is the earth current calculated for the analyzed MV island.

5. Discussion

The current rapid development of prosumer RESs cause various problems in distribution networks. One of them is the difficulty in carrying out planned maintenance in LV and MV networks. The high saturation of RESs in LV networks may result in active power flow towards the MV network. This often makes it impossible to traditionally feed a section of the MV network from a diesel generator as this leads to generator shutdown due to reverse power. One solution is to use a hybrid mobile station consisting of a diesel generator, energy storage, and an LV/MV transformer. Such an arrangement offers the possibility of actively balancing power in a dedicated MV island. HMS capacity is determined by the load profile, which can be calculated based on historical measurements and weather forecasts.

The second problem associated with powering the MV island using the HMS is verification that electric shock protection requirements are met. The need for this assessment arises from the change in operation of the MV network from compensated to isolated. This can be calculated using the network parameters of the islanded MV network section.

Therefore, in addition to the HMS, computation software is needed to determine the HMS rating from the load profiles and to assess the electric shock protection. A simplified algorithm for such software is shown in Figure 16. The algorithm starts by indicating the day of scheduled work and checking whether historical measurements are available for the same period as the indicated day. If there are no historical measurements, then another day should be selected. Next, the electric shock protection ability of the HMS and islanded grid is considered. The earth short-circuit current I”_k1island in the considered island is calculated. When the current I^”_k1island is higher than the residual current I_E of the full MV grid then the HMS protection settings can be set, for instance, as equal in protection relay in a 110 kV/MV substation. In another case the HMS protection settings have to be modified.

Next, the HMS capacity sizing process, marked by letter A, can be realized. The process details are presented in Figure 17. Here, the following data are necessary: validated historical load profiles of MV/LV substations; capacity of RESs installed in considered grid now and in the past; weather forecast for day of planned work; and historical weather forecast. This allows us to determine the load profile of the HMS, the required rated power of the diesel generator, and the required energy storage capacity.

The aim of the HMS is to balance the MV island’s power. The power balance variation results from variations in customer load and RES generation. These variations can be significant over short-term periods. The HMS operation basic algorithm is shown in Figure 18. Active power balancing is realized by diesel generator (DG) when the island behaves as a load, i.e., P_HMS > 0, which means power flow to grid. When energy S_ES stored in the energy storage is less than 20%, the DG simultaneously charges the energy storage. In the opposite case, the energy storage operates with power equal to P_ES = 0. When the island behaves as a source of energy, i.e., P_HMS ≤ 0, which means power flow from grid to the HMS, the energy storage ES is responsible for power balancing, while the diesel generator DG operates with power equal to P_DG = 0.

The algorithms presented above are a starting point for developing detailed algorithms.

Nevertheless, such a solution would undoubtedly support distribution network operators, speeding up the planning of maintenance works while increasing power supply reliability and ensuring safety.

The solution described above assumes that the HMS consists of a diesel generator, energy storage, and an LV/MV transformer. Such a combination seems to be optimal in terms of the flexibility of its application. In this case, it is possible to supply the MV island with an uncharged energy storage and balance the power with the energy storage or diesel generator.

On the other hand, when disconnecting the transformer from the HMS it is possible to supply the LV grid with RESs. Nevertheless, another solution could be an HMS consisting only of energy storage and a transformer. But such a solution requires a more precise determination of the HMS load profile and assumed work time.

The proposed HMS is a solution for grids with an increasing capacity of RESs. The energy storage and MV/LV transformer increase the cost of such an HMS but also increase its functionality, significantly simplify logistics and saving human and material resources. For this reason, the implementation of this type of solution in practice should be expected.

Another advantage of the proposed system is the HMS mobility. This is very important for DSOs, because in practice, the HMS is connected to an MV tower which is sometimes in terrain with difficult access. Thus, a mobile solution, i.e., built on a car, is necessary. The mobility feature unfortunately limits the energy storage capacity. This leads to the state that it is not possible to use such an HMS in every case (day).