Supercapacitors are in many ways better than traditional batteries for energy storage:
- they can (usually) be charged much faster than batteries,
- they have longer lifespan (including more charge/discharge cycles),
- they have broader usable temperature range (batteries tend to lose charge in low temperatures),
- they can (usually) provide higher discharge current than batteries
but as they are usually rated for maximum voltage not exceeding 3 V, there are some implementation challenges.
As typical capacitor can (unlike the battery which can be quickly destroyed by that kind of operation) be safely discharged all the way to 0 V, getting all the stored energy can require special converter designs, capable of working with near 0 V input voltage.
Another approach - known from standard battery construction - is serial connecting several supercapacitors to get storage bank with higher working voltage, but then the cell balancing problem appears. It can happen in three stages of storage bank operation:
- voltage imbalance during charge phase (caused by capacitance difference between individual capacitors),
- voltage imbalance during energy storage (caused by different leakage currents),
- voltage imbalance during discharge phase (caused by the difference of currently stored charge between cells)
Those processes can lead to one of two adverse effects - either exceeding maximum voltage of some cells (during charge phase) or - during discharge -premature depleting some cells and in effect subsequently charging them with inversed polarity.
Supercapacitor balancing techniques can be divided into three groups:
- none/external (when capacitor banks are constructed using capacitors of known capacitance value/tolerance/leakage current, then operated within safe parameter margins),
- passive (when individual cells are protected by shunts or elements of regulated resistance),
- and active when cells are managed using active switching and/or amplifying circuits
To better understand supercapacitor bank construction challenges I plan to:
- perform some testing using capacitor theoretical model (in which supercapacitor is modelled as parallel network of individual sub-capacitors, each connected through different series resistance) - I hope to be able to estimate how much of the capacitance is connected through resistances higher than nominal (that would be beneficial for deciding about effect of fast/pulse charging),
- research about present techniques used in balancing series connected supercapacitor banks,
- as I am in the possession of some very old 5V rated supercapacitors (or - to be more precise, two capacitor series modules) I would like to test how the old age influenced their internal balance,
- build some passive BMS circuits and compare them,
- research data about hybrid supercapacitors and try to compare them to standard models
To conduct those experiments I would like to use component set generously provided to us by our Sponsors, Cornell Dubilier and Element14, consisting of (from the official list and the invoice, because my challenger's set is currently waiting for the custom's clearance):
- three sets of two capacitors rated for 2.7V and with capacitance of 1F, 5F, 10F,
- three sets of two capacitors rated for 3V and with capacitance of 3F, 7F, 25F,
- set of two capacitors rated for 3.6V and with capacitance of 0.22F,
- two/three sets of two hybrid capacitors rated for 3.8V and with capacitance of 25F, 40F,
- three sets of two capacitors rated for 5.5V and with capacitance of 0.1F, 0.47F, 1F,
and some supercapacitors currently in my possession.
