Decentralised storage: Factors to consider in designing the perfect residential system

SMA's vision of a connected home, complete with storage and solar PV. Image: SMA.
In our last blog post for PV Tech Storage, we discussed the boundary conditions influencing the deployment of end customer owned batteries. In this blog article we will share some experiences from technical discussions over the past few years.

Battery sizing – driven by economics or emotion?

The topic which dominated the discussion at the very beginning was the sizing of the storage needed. An early approach was “the more, the better”, somehow also linked to some more or less emotional if not irrational reasons, e.g. the wish to be able to cover the complete energy demand over the whole night out of the battery. Although there is economically no point for this “night-reason”, it seemed to be common sense among spontaneous customer reactions.

However, if the size of the battery is related to the PV profile and load profile, sooner or later it started becoming obvious that a big battery can only be fully charged on very few days a year, but- and this is bad news- cannot be completely discharged on these very sunny days. Oversizing the battery for emotional reasons means that you will not use all of the storage capacity but will instead have “dead capital” in your basement.

Is there something like the “right technology”?

The question of selecting the “best” storage technology was prominent in early discussions. Lead-acid based systems, the first ones available on the market, from the very beginning had to fight an image of providing an old-fashioned, if not boring, technology. At the same time, Li-ion (LIB), as the only “hip” technology entered the market providing apparently higher benefits than its competitors.

The share of LIB based systems in the market is steadily increasing- and this is good news- accompanied by tremendous cost reduction potentials. Datasheets showing much higher cell-efficiencies, much wider utilisable depth of discharge (DOD) ranges and much better cycle performances catapulted the LIB to being the “battery of choice”. Today, we see that this technology is showing good results in the field, costs are reaching a very competitive level, so the answer looks quite obvious – using LIB technology leads us in the right direction.

But is choosing the right battery technology the only thing you have to do? To be honest, the cell performance does not tell the whole truth about the battery: as LIB cells need more “care” than lead-acid cells, mainly for safety reasons, Li-systems have to be monitored accurately and thus will be more complex. Due to the internal consumption of the electronic components and safety elements included in the battery system, the overall battery efficiency can even be worse than the efficiency of a state-of-the-art lead-acid system.

The wider DOD range and cycle performance are relevant aspects only if you can utilise them. Oversizing the LIB system will result in small DOD as explained above - and due to the larger size the number of possible cycles shown in the data sheet will never be reached within the service life of the battery. In order to really utilize the advantages of a LIB battery, it should use up the higher available cycle number and thus be comparably small - or survive approximately 20 years. This is one of the reasons why today, compared to five years ago, you will find many batteries claiming 20 years of calendric lifetime – up until now, there is no proof in the field that this is possible, only some laboratory tests. However, the performance of the overall system is not only dependent on the chosen technology but more on intelligent system design and sizing instead.

Lithium-ion cells. Design of home storage is increasingly about looking at the whole system, not just the choice of battery or coupling method, according to Volker Wachenfeld and Aleksandra-Sasa Bukvic-Schaefer.
What makes a well-designed system?

This leads us to the next controversial topic, according to our discussions. The initial discussion about a “well-designed system” was reduced to one simple point of differentiation: should the battery be coupled on the DC or on the AC side of the PV system?

The basic philosophy behind DC coupling is to combine the PV generator and storage together on a DC link via DC/DC converters and use just one DC/AC inverter as grid interface. In contrast to that, AC coupling requires separate inverters for the PV generator and for the battery. As only one inverter is sufficient for DC coupling, it sounds reasonable that DC coupled systems should be more efficient and more cost effective. Looking into detail, this general finding does not survive a closer examination.

If the overall design of a DC coupled system is performed properly in any dimension, we can say yes, that’s the best way to do it. But, in real life, not many of the systems in the market will meet this basic requirement. Comparing measurement results of systems offered on the market today, you can see some well-designed AC coupled systems showing better characteristics than most of the DC systems. Only the highly integrated DC systems really show advantages, especially in terms of efficiency.

Besides this, a “well-designed system” also depends on the share of the PV energy that is to be stored compared to the share of PV energy that is directly used - and this depends mainly on the degree of matching of load and generation. For example, if a larger share of the generated power can be consumed directly, the AC coupled system can prove a better performance ratio in annual average.

One of the major arguments for AC coupled systems is their ease of use in retrofit applications, upgrading existing PV plants with battery storage. In this case, the existing inverter can be used, whereas installing a DC coupled system would require an exchange of the original PV inverter. Furthermore, AC coupled systems usually provide better performance for backup applications. In the end, a “well-designed system” cannot be reduced to the preferred way of coupling the battery – all aspects quoted here and more have to be taken into account.

"…the success of storage will be directly connected to the lowest possible costs for storing energy." Image: SMA.
So, what is the right system approach?

In the end, the decisive factor should be the “levelised cost of stored energy”: we could call that LCOS(E). No matter if it is an AC or DC coupled system, if LIB or lead-acid batteries are used, the success of storage will be directly connected to the lowest possible costs for storing energy. And the lowest costs will essentially be a question of the right system approach.

The system borders we have to analyse include battery cells, modules, the BMS, power electronics and their interfaces as well as sensors and algorithms. Starting with the design, PV generator and battery sizing have to be well synchronised – any kind of mismatch causes reduced efficiency and thus unnecessary costs. The overall performance ratio can be optimised by utilization of flexible loads – generally speaking, load shifting via demand side management is more efficient than storing energy in a battery.

Even integrating thermal storage could add value to the system’s performance ratio. To allow an intelligent charging strategy, predictive data for generation from weather forecasts and load demand, locally measured and calculated, have to be taken into account. The better the system “understands” the local load profile, the more intelligent operational strategies for the battery can be implemented. And last but not least: an open interface for any kind of external control system, such as a supervisory control and data acquisition (SCADA) system of the local DSO, will allow for operation of the battery in multivalent use cases and thus make it “future proof”.

Let me finish with one more aspect that has to be taken into consideration when designing a residential storage system: it will not be possible to do any kind of specific engineering on a household basis. Costs for consulting, procurement and specific installation requirements will not allow for any kind of customising.

We tried to point out that the best possible storage sizing has to follow the specific load profile, the expected maximum load and the PV generation profile. Nevertheless, to offer a storage system on a mass market, finding the best compromise is mandatory. But that should be possible – on most markets, typical load and generation profiles are available.

This article was written collaboratively by Dr. Alexsandra Sasa Bukvic-Schaefer in collaboration with her colleague, Volker Wachenfeld, senior vice president for hybrid and energy storage integration at SMA Solar Technology.