A hybrid energy storage system pairs a supercapacitor with a battery, and the reason is not that one is better than the other. It is that they cover different time scales. A battery stores energy chemically and moves it slowly; a supercapacitor stores it as charge and moves it in fractions of a second. Put both behind the same bus and each one handles the part of the duty cycle it is built for. What decides whether the arrangement works is where you draw the line between them, and how the control system enforces it.
This article opens a series on hybrid storage. It deals with the split itself: what belongs to the capacitor, what belongs to the battery, how the two are coupled, and what goes wrong when the boundary is drawn in the wrong place.
The Split Is by Time Scale, Not by Capacity
Dividing the work by size, letting the big bank carry the bulk and the small one cover the peaks, leads to the wrong specification. A supercapacitor bank that looks small in energy terms can be enormous in power terms, and the opposite is true of a battery bank.
The useful axis is time. Ask how long a given demand lasts, and the answer usually falls on one side of a line somewhere between a few seconds and a few minutes.
| Duration of the demand | What it looks like in practice | Device that fits |
|---|---|---|
| Milliseconds to a few seconds | Switching transients, motor starts, fault ride-through, frequency response | Supercapacitor |
| Seconds to a few minutes | Ramp smoothing, peak shaving, bridge power during a generator start | Either, depending on how often it repeats |
| Minutes to hours | Load shifting, time-of-use arbitrage, scheduled discharge | Battery |
The middle row is where designs get argued about, and the deciding factor is how many times per day that duration occurs. A two-minute event that happens twice a day is battery territory. The same two-minute event happening two hundred times a day is not.
What High-Frequency Cycling Costs a Battery
Batteries fail in two ways: they wear out with age, and they wear out with cycles. The second is what a hybrid system is buying protection against.
Every charge and discharge moves lithium between electrodes through a process that is not perfectly reversible. Small amounts of material are consumed at each cycle, and the loss accumulates. A cell rated for several thousand cycles at moderate depth will not deliver that count if it spends its life absorbing short, deep pulses, because each pulse still counts as a cycle even when the energy moved is trivial.
This is why the economics of hybrid storage are usually argued in cycles rather than in watts. What a hybrid system protects the battery from is a job measured in millions of small events rather than thousands of large ones. The overload was never the problem.
The Slice a Supercapacitor Covers
A supercapacitor covers the short end of the duty cycle, and it does so without a chemical reaction at the electrode, which is why its cycle life is quoted in the hundreds of thousands rather than the thousands.
Frequency response falls into that slice most cleanly: the demand is a burst of power lasting a few seconds and repeating constantly, which suits a capacitor almost perfectly. Peak shaving on a load with sharp, frequent spikes is the second. The third is ride-through, where the system has to hold up through a voltage dip or a generator start long enough for something else to take over, a window usually measured in seconds.
Station cells are the building block for these duties. A 2.7 V cell in the tens to hundreds of farads covers small systems directly; where the bus voltage is higher, cells stack into modules, and the module’s usable energy is set by how far the voltage is allowed to swing rather than by the nameplate value.
Two Ways to Couple Them
The battery and the capacitor can sit on the same DC bus, or they can be separated by a converter. The choice drives the cost, the efficiency and the control complexity at the same time.
| Aspect | Direct parallel on the DC bus | Separated by a DC-DC converter |
|---|---|---|
| Voltage | The capacitor bank has to be built to the bus voltage, so it takes more cells in series | The capacitor can run at its own voltage, so fewer cells in series |
| Usable energy | Limited: the capacitor only contributes over the voltage range the bus allows | Higher: the converter can swing the capacitor across its full range |
| Losses | None beyond the capacitor’s own ESR | Converter losses on every transfer |
| Control | Passive; the split happens by itself through impedance | Active; the converter decides what flows where |
Direct coupling is the simpler and often the better choice when the capacitor only has to cover seconds. Its weakness is that the split is then decided by impedance rather than by intent: the capacitor takes the fast current because it presents the lower impedance at high frequency, and how well that works depends on the wiring as much as on the components.
Converter coupling earns its cost when the capacitor has to contribute a meaningful share of the energy, because it lets the capacitor voltage swing over a wider range and therefore extract more of what the cell holds. The trade is a conversion loss on every event, which matters if the events are frequent.
The Control Question Nobody Budgets For
In a converter-coupled system, the allocation is written in software, and that is where hybrid projects fail quietly. A controller that gives the battery a slow response and the capacitor nothing to do is not a hybrid system; it is a battery system with an expensive capacitor attached.
The usual arrangement is a high-pass and low-pass split: the capacitor absorbs the fast component of the demand, the battery follows the slow one. Getting this wrong is not obvious from the outside, because the system still meets its energy requirement. What degrades is the reason the capacitor was added: the battery keeps seeing the fast cycling, and the failure simply arrives later than it otherwise would have.
Sizing the Split
The capacitor’s size follows from the energy in the events it is meant to absorb, not from a percentage of the battery’s capacity. Size it from the worst short event: the power it must deliver, the duration it must sustain, and the voltage swing the coupling allows. Then check the resulting current against the ESR and the thermal limit of the cell or module.
Check the answer against the cycle count as well. If the events are frequent enough that the battery’s cycle budget would be exhausted within its calendar life, the capacitor is doing real work. If they are not, the capacitor may still be justified for ride-through or for power quality, but the lifetime argument for it is weaker.
Where the Series Goes Next
Setting the split is the calculation. Making it hold on a real DC bus is the harder part, and it comes down to how the converter is rated and how the allocation is tuned once the system is running rather than worked out on paper. That is where the next article in this series goes.
Working out which device suits a given load on its own is a separate question, and the load profile comparison answers it. If you need the wider view of what is available before choosing, start from the selection guide; the cells and modules used to build a hybrid bank are in the module range.
Common Questions About Hybrid Energy Storage
Do I still need the battery if I add a supercapacitor?
Yes. The capacitor covers seconds; the battery covers everything longer. A capacitor bank large enough to replace the battery’s energy would be impractical in both size and cost, so the two are complementary rather than alternatives.
How do I know if my load actually needs a hybrid system?
Look at the event count rather than the event size. If the short, high-power events are frequent enough that the battery’s cycle budget would be consumed well inside its service life, the hybrid arrangement pays for itself. If they are occasional, a battery alone will usually be cheaper.
Can I just parallel a supercapacitor with my existing battery bank?
Electrically yes, and it will take the fast current because its impedance is lower at high frequency. What you cannot do that way is decide how much of the energy comes from each device, because the split is set by impedance and wiring rather than by design. If the capacitor only needs to cover transients, that may be enough.
What voltage should the capacitor bank be built for?
Either the bus voltage, if it is directly paralleled, or its own lower voltage if a converter sits between them. A direct-coupled bank needs more cells in series, which also means more cells to balance.
Xuansn manufactures supercapacitor cells and modules across the 2.7 V, 3.0 V and 5.5 V families, including the 3.8 V lithium-ion capacitor range and the energy storage modules used in hybrid banks. Send us the power profile of the events you need to cover and we will work out the split with you.
✉ coco@xuanxcapacitors.com | ☎ +86-769-8166 8821 | Send your power profile
