2.1 Energy Storage for PV

In Germany, the vast majority of the installed PV capacity, about 13 GW, is connected to the LV grid. Situations of overvoltage are visible throughout the year in different areas. To limit such events, since January 2012, a fixed limitation of the active power feed-in by each PV system has been mandatory [6]. At the time of writing, the limit is set to the 70% of the maximum power of a PV system; this corresponds to indirectly curtail up to 30% of the available power during periods of maximum PV generation. Alternative solutions to power curtailment may involve an increased local consumption rate that can be achieved by domestic load and energy storage.

Local consumption methods in residential networks with PV have mainly converged on demand-side management (DSM) to shape the domestic power demand during periods of high PV generation [7]. However, as the number of flexible domestic appliances is very limited and these may not necessarily be used on a daily basis (e.g., washing machines) local consumption can be significantly aided with the deployment of battery-based energy-storage systems (ESS). ESS could help to solve such problems by storing energy when PV generation is high and demand is low, and releasing it back to the grid when generation is low. Private owners of ESS and PV would have smoother and more predictable load profiles, ensuring a better network operation and allowing the LV grids to host more PV systems. ESS can also be used to facilitate the participation of PV power plants in the electricity market operation, by providing committed power during certain periods of the day or frequency regulation according to grid codes [8].

The conventional size of domestic ESS for private households with PV is in the range of 5 kWh, from which a local consumption rate of about 55% on a 12-month period can be achieved [9]. The conventional way of performing local consumption consists of activating the charging of the battery as soon as the PV-generated power is greater than the domestic power demand. However, this strategy does not guarantee that the voltage limits are not exceeded during the periods of peak PV generation (12:00–2:00 PM), because the battery gets fully charged during the morning hours.

In the near future, a significant portion of new PV installations is likely to be grid-connected to the residential LV grid, coupled with an ESS.

2.2 Energy Storage in feeders with PV

In residential feeders with a high penetration of PV, the PV-power generation sees its maximum in a time interval around noon. The PV integration should be realized by ensuring that the grid components such as the distribution transformer, lines, and cables operate within their physical limits. At the same time, the voltage profiles at the different buses should meet the power-quality requirements, as stated by the power-quality standard EN 50160. Under high-generation and low-load conditions, one of the goals is to limit the risk that the feeder voltages exceed the allowed limits at the different buses. In Figure 10.1, a simplified feeder is used to explain the case of a LV feeder with PV. When the power generated from PV becomes larger than consumption, the voltage magnitude along the feeder increases, creating potentially overvoltage problems.

where

P_PV is the active power generated from PV at the feeder bus

P_L, Q_L is the load at the feeder bus at the feeder bus

ΔU is the voltage rise due to PV generation at the feeder bus

It is clear that the power injected into the feeder by PV should be counterbalanced by an energy-storage system that locally stores the excess energy produced. While this scenario is increasingly becoming an urgent need to preserve power quality in the feeder, private-storage strategies have in fact been developed with a different focus. The original goal for introducing residential storage systems was to increase the local consumption rate of PV energy. This conventional storage strategy prioritizes the charge of the battery as soon as the PV production exceeds the household demand, regardless of the grid power-quality needs [9]. Thus, while it allows a reduction in energy intake from the mains, the conventional strategy does not bring any advantage to the grid operation.

On the other hand, by prioritizing the charge of the battery around peak generation hours, a grid-supportive storage strategy can be enabled. With the latter, a decentralized storage concept can be put in place, so that each ESS is activated at a certain power threshold, P_th, which may be pre-determined in advance for every feeder.

In Figure 10.2, the two described storage strategies are represented. The conventional storage strategy activates battery charging as soon as power generation exceeds the demand. The grid-supportive storage strategy delays the charging activation until the maximum generation period.

The combination of the grid-supportive storage strategy with the conventional one is also possible today by the use of day-ahead solar-irradiation forecasts. This combination would be particularly useful on those days with more unpredictable solar irradiation.

3.1 Grid-supportive energy storage

To avoid situations of overvoltage during high-PV generation periods, the ESS at each household with PV should be activated during the maximum-generation interval, as illustrated in Figure 10.2. This can be achieved by deploying a house main controller which automatically reacts based on the actual PV generation. The house main controller controls the activation of the ESS according to a pre-defined power threshold P_th. If the PV power reaches the threshold P_th, e.g. 70% of the PV maximum power, the main controller activates the charging of the ESS battery. At the end of the PV generation period, after sunset, the main controller stops the charge and sets the ESS to perform load shading for the domestic appliances. Thus, depending on the time of day, the main controller sets the battery to charge, discharge, or idle mode. During the charging phase, the ESS battery stores the incoming PV energy with a power level P_S = P_PV – P_th. The charging process continues as long as P_PV is greater than the preset threshold.

The safe control of the ESS battery should be ensured at all times, during charging and discharging operation. This task is achieved on the ESS state-of-charge (SOC) control, which is the innermost control loop of the storage system. The charging of the ESS battery should be limited to a safe SOC window, which for a lithium-ion battery is considered to be within 20 to 90% [10]. Other types of batteries may also be considered, requiring potentially a different usage window.

To provide voltage support in the feeder, the maximum charging power of the ESS under worst-case conditions will vary according to the preset power threshold. For LV grids with different PV-penetration levels, this power threshold can be identified according to the grid configuration.

4.1 A real example

The Belgian residential LV feeder presented in [11] is an interesting example of a LV grid with high penetration of PV and overvoltage issues. The feeder is composed of seven buses, with decentralized PV systems at different feeder locations as shown in Figure 10.4.

The PV installed capacity at the different buses in the feeder is distributed as indicated in Table 10.1.

The LV feeder is part of a larger LV grid, which includes nine feeders. The feeder supplies 33 houses and it embeds nine roof-mounted PV systems, with total PV installed capacity of 42.6 kW. On buses 2, 4, and 5, the size of each PV system corresponds to the value indicated in Table 10.1. The definition of PV penetration is given as follows:

${\mathrm{PV}}_{\mathrm{penetration}}=\frac{\mathrm{PV}\mathrm{installed}\mathrm{capacity}}{\mathrm{feeder}\mathrm{capacity}}\text{.}$

The PV penetration in the feeder is the ratio between the total PV installed capacity to the nominal feeder capacity. The feeder capacity is intended as the capacity of the first-line section of the feeder, corresponding to 185 kVA. It follows that a PV penetration of 23% is present in the Belgian LV feeder.

The measured profiles of load and generation in the feeder are represented from January 1st until December 31st in Figures 10.5 (a) and (b), respectively. The same profiles are used in the simulation model to evaluate the situations of overvoltage. The 12-month simulation performed shows that overvoltage events at bus 7 are more critical, and the maximum voltage limit of 1.1 p.u. is exceeded in 98 days.

In Figure 10.5 (c), the difference “PV generation – load” in those days with voltages exceeding 1.1 p.u. is depicted. Some major findings are:

1) overvoltage events occur when G-L is positive; in the field case, this corresponds to the active power range of 15 to 35 kW;

2) the highest concentration of days with voltage above the 1.1 p.u. is observed in spring and summer.

With regard to grid losses, the total energy losses in the feeder are calculated considering the losses on the single-feeder cables; the profile of total power losses in the feeder during the 12-month period is depicted in Figure 10.6. Losses are lower in spring and summer and higher in winter and autumn.

With regard to energy levels, the total energy consumption, energy production, and energy losses are calculated and summarized in Table 10.2.

A worst-case scenario for overvoltage in the feeder is observed during sunny days of spring and summer, according to the measurements, and during low load conditions.

If by hypothesis 50% PV penetration is assumed, the PV installed capacity per bus would result in the one indicated in Table 10.3. With a 50% penetration level, the optimization method identifies a power threshold value of 45% of the rated PV system power capacity for each house, which corresponds to a higher charging power for each ESS in the feeder. By doubling the PV penetration in the feeder, the ESS power requirement becomes much higher. Consequently, the ESS energy requirement will be quite different in the two scenarios.

5.1 Case 1

Under Case 1, the feed-in power by the PV system is at most the 80% of P_PVmax. The charging activity of the nine ESS is depicted in Figure 10.7. The maximum charging power is about 1.1 kW, which is in line with the LP method’s results. The phase voltage profiles Va, Vb, and Vc of the most critical feeder location, bus 7, are depicted in Figure 10.8. The profiles are apparently within the maximum voltage limit.

5.2 Case 2

With the power threshold of the PV system set to the 70% of P_PVmax, the charging activity of the residential ESSs is depicted in Figure 10.9. In this case, the maximum charging power is about 1.7 kW and all ESS operate for 147 days. The phase-voltage profiles Va, Vb, and Vc of bus 7 are depicted in Figure 10.10. Compared to the voltage profiles obtained for P_th = 80%, a wider margin from the limit of 1.1 p.u. is observable.

5.3 Case 3

With 50% PV penetration, the aggregated generated power by PV in the feeder is up to 92 kW. Such power amount leads to 231 days with overvoltage situations, which would be apparently an unacceptable power-quality scenario. A decentralized storage strategy requires in this case a power threshold of 50%, to avoid overvoltage and to comply with the power-quality standard.

5.4 Storage Energy Levels

The ESS energy levels can be obtained for the power profiles obtained for the three cases.

With 23% PV penetration and P_th = 80%, the required ESS energy is 1.1 kWh, as seen in Figure 10.11. This value is obtained by integration over time of the ESS power profiles and selection of the day with the highest storage-energy content, whereas, with P_th = 70%, the required ESS energy is 3.8 kWh, as seen in Figure 10.12.

With 50% PV penetration and P_th = 50%, the required ESS energy capacity with nine ESSs in the feeder is 28 kWh per unit, whereas it is 14 kWh with 18 ESSs.

It should be considered that the identified energy levels are referred to as the operative energy window (in kWh) for achieving voltage support using decentralized ESSs. To operate the ESS battery across a linear SOC region and to avoid deep discharge cycles, the battery should be oversized in relation to the operative energy window identified. For each ESS battery, a usable SOC window of 20 to 90% is imposed in this chapter, complying with common recommendations on the use of Li-ion batteries in [10]. Following the nominal battery capacity E_N, the minimum and maximum energy constraints E_min and E_max are indicated in Table 10.4.

Table 10.4

ESS Operative Constraints

Case	Nr. PV systems	EN (kWh)	Emin (kWh)	Emax (kWh)	${{\mathit{P}}_c}^{\mathbf{max}}$ (kW)	${{\mathit{P}}_d}^{\mathbf{max}}$ (kW)
PV penetration: 23%
Pth = 80%	9	1.6	0.3	1.4	2	-2
Pth = 70%		5.4	1.1	4.9	2	-2
PV penetration: 50%
Pth = 50%	9	40	8	36	6	-6

With 23% PV penetration, voltage support with P_th = 80% is possible with a nominal battery capacity of about 1.6 kWh. Instead, with P_th = 70%, a battery capacity of 5.4 kWh should be used. While a smaller battery size represents a more affordable option for the user, the larger battery size gives the opportunity to achieve a higher local consumption rate. It is also evident that the maximum charging and discharging power of the ESS is not a crucial issue for the provision of voltage support; in fact, the power level of ± 2 kW is enough for both cases with 23% penetration.

Drastic differences are observed in power and energy levels with 50% PV penetration. With such penetration level, the required ESS nominal capacity is not an economically feasible option for standard households.

6.1 Energy losses in the feeder

It is also worth looking at power losses in the feeder with and without the decentralized energy storage strategy applied.

A comparison of energy losses in the feeder with 23% PV penetration is performed for the two cases of decentralized storage with 70% and 80% power threshold and for the case without storage. With the 80% power threshold, the losses are reduced by 6.6% compared to the case without storage. If decentralized storage is activated at 70% power threshold, losses are reduced to about 7.3%.