Large scale storage for utility scale solar PV


With the increase in very large utility solar photovoltaic (PV) installations, it is becoming common practice to install utility scale storage together with PV solar farms, rather than rely on grid storage or user storage. This is influenced by the choice to rely on controllable output than external factors such as grid interconnections. The addition of storage adds stability and controllability to large scale PV plant, but adds cost and technical complexity. Design and sizing of storage depend on the application envisaged.

As the penetration of renewables increases, more demands are being placed on reliability and dispatchability by network operators. The initial priority dispatch arrangement where there was no control over the output of a solar farm and all output was dispatched is proving to be problematic, and operators are aiming at providing high quality power when needed. In addition pricing systems are moving from a guaranteed tariff to market or demand driven principles. Some systems offer a premium for power during peak demand periods, and levy penalties for variable supply.

Developments in battery energy storage technologies have led to decreasing prices and increase in size, with the result that an increasing number of large storage units are being installed in networks worldwide. Up to now the focus has primarily been on network storage for large capacity plant, but there is an increasing trend to install storage together with solar PV on site. Systems with an on-site capacity of up to 30% of the total PV capacity are appearing. An announcement has been made of a 300 MW solar farm with 100 MW storage and 400 MWh capacity [1].

Table 1: Storage required to achieve ramp rate constraint requirements, for a 1,6 MW system [5].
Maximum allowed ramp rate Battery power for 100% compliance Battery power for 99,5 % compliance Battery power for 98% compliance
100%/10 min 0 0 0
20%/2 min 47% 37% 14%
10%/1 min 55% 45% 23%
5%/30 sec 59% 48% 26%
1%/6 sec 60% 50% 27%
0,16%/1sec 64% 52% 27%

In addition there are reports of planned systems comprising 120 MW of solar with 100 MW/200 MWh storage [1]. The companies involved claim that large-scale renewables and large-scale battery technology will play a central role in keeping the electricity system stable, reducing prices, and reducing emissions [1].

There are no plans for large scale battery storage coupled with PV in the South African utility market, and neither have any of the several proposed plans or studies for the development of the electricity generation network seriously considered storage as a component. The only coupled storage systems are those associated with CSP plant. There are a number of small scale trial non-thermal storage systems in operation. This may change in future as electricity storage becomes cheaper and the renewable energy network develops, and it will useful to consider some of the factors which affect the design of large scale electricity storage systems coupled with solar farms.

PV and storage parameters

PV sizing requirements

The PV array must have the capacity to supply the agreed dispatchable power and recharge the storage within the required time. The storage charging portion will be required to recharge the storage after a discharge, in addition to delivering power to the grid. The charging period will depend on the depth of discharge and the interval between discharges

Take the example above of 300 MW with 100 MWh storage. A 300 MW plant will produce 300 × 6 = 1800 MWh of energy per day. Taking 100 MWh required for storage, the unit will only be able to deliver 1700 MWh/day and have an equivalent capacity of 283 MW which can be delivered to the grid.

Storage sizing requirements

The storage size has two components, output power (MW) and energy (MWh). The two are related but not interdependent. A battery may be required to deliver high power for short periods, giving a low energy requirement, or at the other extreme, low power for a long period, giving a high energy requirement but low power output. The battery sizing will be given in terms of both, i.e. MW output and MWh capacity (e.g. 100 MW/50 MWh).

Fig. 1: Ramp rate constraints.

Requirements for storage

One challenge for renewable power project developers relates to the efficient design of storage capacity to cost-effectively meet the grid utility’s requirements. To accurately size the electricity storage system (ESS), battery sizing needs to be based on high-resolution data (i.e. one minute or less) of the renewable resource or power output at the site location. In contrast, it is currently not general practice for solar PV plants to have an on-site weather station before the plant enters operation. It is unknown whether such information is gathered for South African sites. Even for countries with abundant national meteorological stations, irradiation data from such public data sources is mostly available at only an hourly, daily or monthly resolution.

Storage can be provided to meet several different requirements. The size and configuration will depend on the requirements. Several typical applications are listed below:

  • Meeting grid code requirements in terms of ramp rate (RR).
  • Smoothing out short term variations in output due to passing clouds.
  • Providing additional dispatchable power outside of the solar radiation window. i.e. extending the dispatchable operation of the solar farm.

Grid code requirements

Short-term variability in the power generated by large grid-connected PV plants can negatively affect power quality and the network reliability. New grid-codes require combining the PV generator with some form of energy storage technology in order to reduce short-term PV power fluctuation. Many grid codes limit the RR rate or gradient rate of power injected into the grid. RR limits range from 1%/min to 10%/min.

The South African grid code specifies a power gradient constraint but does not stipulate a maximum ramp rate [2]. A power gradient constraint is used to limit the maximum ramp rates by which the active power can be changed in the event of changes in primary renewable energy supply or the setpoints for the solar plant, taking into account the availability of primary energy to support these gradients.

A power gradient constraint is typically used for reasons of system operation to prevent changes in active power from impacting the stability of the transmission system or distribution system. The solar plant control system is required to be capable of controlling the RR of its active power output with a maximum MW/min ramp rate set by the system operator. These RR settings are applicable for all ranges of operation including positive ramp rate during start up, positive ramp rate only during normal operation and negative ramp rate during controlled shut down.

Fig. 2: Ramp rate control method [4].

No information is given in the grid code on the RR to be used in the South African system and it is not known whether the power gradient constraint is active in the current network [2]. Power gradient constraints are applied in other networks, with penalties or disconnection in the case of non-compliance. Disconnection results in curtailment costs, i.e. the value of lost production.

Ramping up and down can occur as a result of changes in the primary energy source, i.e. the solar resource. Cloud cover moving over a solar farm can result in very sharp changes in power output. Experience with several solar facilities in the US shows that passing clouds can cause the output of PV resources to drop from 100% output to 20% in less than one minute and to return to 100% just as quickly [7].

Other studies have shown that fluctuation depends on the size of the solar farm, and the larger the PV system is, the lower the PV fluctuations are: typically, for a 1 min time window, a 1 MW PV plant records fluctuations of up to 90%, a 9,5 MW plant easily exceeds 70%, while a 45,6 MW PV plant fluctuates by up to 33% [8].

There are a variety of different methods used for calculating the storage required to achieve compliance with the grid code. All methods make use of analysing short interval records of solar radiation on the site. Less accurate results could be achieved by using one site as representative for an area.

Table 1 shows the storage requirements for a 1,6 MW PV system [5]. The single ramp rate 10%/min is expressed several ways in the first column. Battery power is expressed in AC power, as a fraction of PV nameplate DC power. Ramp rate smoothing only operates when the allowable ramp rate is exceeded, thus limiting the storage capacity requirements of the battery.

One of the simpler methods is straightforward ramp rate control. The modelled PV power is fed through the RR limiter to calculate the desired limit compliant grid feed-in power. The limiter simply forces the difference between the present and the next value of to be within a specified RR limit.

The amount of ESS power can then be determined as the difference between the desired grid feed-in power and the PV power. In this manner, the fluctuations of output are compensated by charging the ESS with the excess power during the upward ramps and discharging the ESS during the downward ramps. Ramp rate control is illustrated in Fig. 2.

Fig. 3: Smoothing of short interruptions [6].

Other methods include the moving average control system, and a step rate control system [4].

G(t) = solar resource variation
Ppv(t) = solar array output
Pg(t) = power output to the grid
Pbat(t) = power supplied by the battery

The power rating of the storage battery can be significant, depending on the solar radiation pattern. Some studies showed values as high a 50% of the PV rating. Although the power requirements are high, the energy requirement is not as the system only operates over a short period. All of the above methods are based on achieving full compliance with grid code but do not measure the cost of achieving this goal. A study by Mott McDonald uses the cost of curtailment or fines as a basis for comparing various battery sizes, i.e. the user accepts a certain amount of losses and minimises the total cost of mitigation and losses [3].

Smoothing out short term variations

In addition to controlling ramp rates, operators may require the ability to smooth out variations in output power due to short term interruptions in primary solar energy. Interruptions could last for several minutes are repeated interruptions will result in a reduction of the total energy and average power supplied during the day. It is obvious that it is desirable to provide a smooth output which will not cause constant changes on the network.

Fig. 3 shows the operation of such a scheme.

The difference between this system and ramp rate control is that the storage system operates continuously to smooth out variations and not only during high ramp rate periods.

Storage can be used to provide smoothing of small interruptions, although the storage capacity required will be more than that required for ramp rate control. Provision has to be made for charging and recharging the storage and the average output will be lower than the full capacity of the array.

Extension of operation beyond solar day

Currently the only systems in South Africa that use this principle are concentrated solar power (CSP) with heat storage, but advances in battery technology and dropping prices may see battery storage used for this purpose in future in this country. This approach can incorporate ramp rate control and short term smoothing as well.

Fig. 4: Solar radiation pattern for a South African site.

The fundamentals are very simple:

  • The solar array has to be sized to collect the full energy requirement for the operating period of the system, over the solar day. This implies, in simple terms, that if six hours of peak equivalent solar is available, and four additional hours of operation at full load are required, the array must be sized to approximately 1,6 times the peak requirement of the system.
    For example, if 100 MW is the required load the array must be sized at 166 MW.
  • The storage unit will need to have sufficient capacity to store the required energy for the required period. The storage unit will also need to be capable of accepting charge at the rate delivered by the array. For cases where the additional operational period is longer than the collection period, the charge rate will be higher than the discharge rate and this needs to be taken into account in the design.

During the design phase things factors such as efficiency and losses will need to be taken into account.

The application of such a system may vary from the simple example given. For instances the system may be required to supplement other sources at a lower rate than the maximum output, which would alter the design parameters.

Storage for more than one day

This approach is unlikely to be adopted for utility scale systems and is generally used for small high reliability systems. The approach is based on the fact that weather patterns are not regular and most systems can experience periods of several consecutive days of poor solar performance. Claims that geographic diversity would eliminate such occurrences in an interconnected network have not been substantiated.

South Africa is an isolated network and cannot rely on power imports to cater for such occasions, and, as recent weather performance shows, there are occasions when most of the country, including Upington, is experiencing cloudy and rainy weather, which persists for several days. In these cases reliance would be placed on conventional generation plant in the network as the cost and complexity of on-site storage would make it impractical.

To understand this consider the requirements, and the nature of inclement weather. Days of poor solar radiation do not occur as single evenly spaced events but as bunches of consecutive days, followed by a few days of good radiation, followed by another cluster of poor days. See Fig. 4.

The system has to cater for two things: storage of energy to make up the deficit over the bad days, and capacity to replace the stored energy in time for the next discharge period.
Consider for example a case where three days of 40% solar radiation is followed by two days of 100% followed by another two days of 40% radiation. The storage unit must cater for 120% of the full capacity of the array, and must be recharged in two days, i.e. at a rate of 60% of the capacity per day. The array will require oversizing by 60% to cater for the recharge requirements.

In other words, the array will be oversized by 60% and the storage requirement, in addition to the normal 24 h storage, will be 120%, so the complete storage will be sized at 220% of the energy capacity of the system. For such a requirement conventional generation systems would make more sense.


[1] S Vorrath: “World’s biggest solar plus battery storage plant ready to build in south Australia”, Reneweconomy, 20 March 2017.
[2] Nersa: “Grid connection code for renewable energy plants connected to the electrcity transmission system or the distribution system in South Africa”, July 2014.
[3] N Jongsuwanwattana, et al: “Storage systems for utility-scale solar and wind”, Mott McDonald.
[4] I de la Parra, et al: “Comparison of control strategies for PV power ramp-rate limitation using energy storage”, 31st European Photovoltaic Solar Energy Conference and Exhibition Hamburg.
[5] D Cormode: “Comparing ramp rates from large and small PV systems, and selection of batteries for ramp rate control”, 39th IEEE Photovoltaic Specialists Conference, 2013.
[6] T Hund: “Grid-tied PV system energy smoothing”, Sandia national laboratories.
[7] Excel Energy: “Solar to battery energy storage project”, Solar to battery fact sheet.
[8] J Marcos, et al: “Control strategies to smooth short-term power fluctuations in large photovoltaic plants using battery storage systems”, Energies 2014.

Send your comments to

The post Large scale storage for utility scale solar PV appeared first on EE Publishers.

Source: EE plublishers

More news