rss
11

ACT launches second phase of battery test centre, early results in

Print Friendly

ITP Renewables (ITP) is today launching a second phase of its independent battery performance testing at the Canberra Institute of Technology.

Supported by a $870,000 grant from ARENA, the purpose of the project is to verify manufacturers’ claims regarding the performance, integration, and installation of residential battery energy storage systems, and to disseminate the results to the public.

The rationale of the testing is that, particularly in remote applications where reliability is critical, energy system designers and end-users have been reluctant to transition to new battery technologies, despite apparent advantages.

In part, this reluctance is due is due to uncertainty about manufacturers’ claims, which often rely on lab-based tests run at a constant temperature and which lack independent verification.

Phase 1 of the testing commenced in August last year with side-by-side testing of six lithium-ion battery packs, an advanced lead-acid battery bank, and a conventional gel lead-acid battery bank.

Phase 2 adds a further eight lithium-ion technologies, as well as a zinc-bromide flow battery and an aqueous hybrid-ion battery.

Phase 1 batteries installed:

  1. CALB (Lithium Iron Phosphate), produced in China
  2. Ecoult UltraFlex (Advanced Lead Acid), produced in the USA
  3. Exide Sonnenschein (Gel Lead Acid), produced in Germany.
  4. Kokam (Nickel Manganese Cobalt), produced in South Korea
  5. LG Chem (Nickel Manganese Cobalt), produced in South Korea
  6. Samsung (Lithium Manganese Oxide / Nickel Manganese Cobalt), produced in South Korea
  7. Sony (Lithium Iron Phosphate), produced in Japan
  8. Tesla (Nickel Manganese Cobalt), produced in Japan/USA

Phase 2 batteries installed:

  1. Alpha Storion (Lithium Iron Phosphate) produced in China
  2. Ampetus Super Pod (Lithium Iron Phosphate) produced in China
  3. Aquion Aspen, (Sodium-ion – also known as a ‘salt water’ battery) produced in the USA.
  4. BYD BBox (Lithium Iron Phosphate) produced in China
  5. GNB Sonnenschein Lithium (Nickel Manganese Cobalt) produced in Germany
  6. LG Chem RESU HV (Nickel Manganese Cobalt), produced in South Korea
  7. Pylontech (Lithium Iron Phosphate), produced in China
  8. Redflow Zcell (Zinc-Bromide flow) designed in Australia and produced in the USA.
  9. Simpliphi (Lithium Iron Phosphate) produced in the USA
  10. Tesla Powerwall 2 AC (Nickel Manganese Cobalt), produced in the USA

Testing takes place in a climate-controlled room (the “Batt Lab”) at the Canberra Institute of Technology.

Within the Batt Lab temperatures are cycled through hot daytime and cool overnight temperatures, similar to what they would be expected to face in real-world conditions.

Testing Procedure

The key objective of the testing is to measure the batteries’ decrease in storage capacity over time and with energy throughput.

As the batteries are cycled they lose the ability to store as much energy as when they are new.

This phenomenon is typically referred to as capacity fade.

To investigate capacity fade, the lithium-ion battery packs are being discharged to a state of charge (SOC) between 5% and 10% (depending on the allowable limits of the particular Battery Management System (BMS), while the gel lead-acid battery bank is being discharged to a 50% SOC (i.e. 50% of the rated capacity used).

The advanced lead-acid battery is being cycled between 30% and 80% SOC. These operating ranges are in line with manufacturers’ recommendations for each technology.

Battery packs are charged over several hours (mimicking daytime charging from PV), followed by a short rest period, then discharging over a few hours (mimicking the late afternoon, early evening period) followed by another short rest period. In total, there are three charge / discharge cycles per day to accelerate their ageing.

Early results

Some initial results are presented below.

These are still too early to be conclusive with only the first six months’ worth of testing collected and analysed. ITP will publish the first year of testing results in August.

Capacity Fade

Figure 1. Measured discharge capacity relative to first-month measured discharge capacity

Figure 1. Measured discharge capacity relative to first-month measured discharge capacity

Capacity fade is evident for some of the battery packs under test, as expected.

However, for others, the trends are not yet discernible owing to the inherent variability in capacity testing results.

In particular, this variability arises because of imprecision in the SOC estimation conducted by the BMS of some packs.

We are conducting a deeper investigation of the atypical Samsung results in particular. The real trends will become clearer as time goes on.

From the data collected thus far, capacity fade is slightly more for the battery with the highest average temperature (the LG Chem pack).

High cell temperatures increase the rate of irreversible side reactions within cells. These reactions consume the “active” materials within the cells, reducing the available capacity.

The LG Chem battery has the highest energy density of all the packs under test, and has no fans or coolant loops to assist with heat dissipation.

While lab temperatures reflect expected real world ambient temperatures, the cycling regime is necessarily much more aggressive.

Hence, the high energy density reduces the ability of the battery pack to dissipate any heat generated as quickly as the other packs under test.

This causes the higher cell temperatures and likely explains the apparent capacity fade rate under the accelerated test conditions.

The two lead-acid technologies show minimal capacity fade thus far but these have undergone fewer equivalent-full cycles owing to the lower allowable depth-of-discharge.

In future reports, when fade is statistically significant, results will be normalised to account for the varying equivalent-full cycles completed.

Efficiency

Figure 2. Measured discharge capacity relative to measured charge capacity

Figure 2. Measured discharge capacity relative to measured charge capacity

Despite the limited data, already it can be seen that lithium-ion out-performs both the advanced and traditional lead-acid battery packs in terms of round-trip efficiency, despite lead-acid efficiency appearing higher than general expectations.

The initial data suggests that efficiency of 90-93% can be expected for either Li-ion NMC or LFP chemistries, and it is possible that a difference in efficiency will be apparent between the different Li-ion chemistries by the conclusion of the trial.

Efficiency data for the Samsung pack requires further validation before publishing, and will be included in the next report.

The advanced lead-acid battery pack (Ecoult) outperforms the conventional lead-acid (GNB) in terms of round-trip efficiency in the data collected thus far.

The ability of the advanced lead-acid to avoid the majority of the conventional lead-acid’s absorption charge phase is likely to be largely responsible for this result.

Oliver Woldring is a Senior Consultant at IT Power (Australia). More detail about the testing methodology and early results, including detail about charge acceptance and battery cell and pack failure is available at: batterytestcentre.com.au/

 

  

Share this:

  • George Michaelson

    Can’t wait for phase two data!

  • Joe

    I would have liked to see Enphase as part of the testing. …the other big names LG, Sonnen, Tesla are in there. Why not The Enphase?

    • wholisticguy

      I am guessing, but perhaps the low inverter capacity in the enphases doesn’t allow the rapid test cycling that the others do.
      Eg a 250w max charge and discharge means a 10 hour cycle time.

      • solarguy

        I’m don’t where you get 10 hour cycle time from, Enphase usable capacity is 1,100wh and the inverter is 270wh. So max charge and discharge is 4.07 hours. For a battery that retails for $2,300, it’s close to useless.

        • wholisticguy

          I looked up the charge cycle protocol on their page, and though my figures might be a bit out, it doesn’t fit their test. The 1140 Wh / 260 W (270VA) give 4.4 hours to discharge, plus 1 hour rest, then 4.4 hours to recharge gives a total cycle time of 9.8 hours. In the testing protocol, each complete charging or discharging period need to be done in 3 hours, not 4.4.

          • solarguy

            So it’s a good bin filler.

  • solarguy

    Lead acid performance isn’t too bad compared to LI-ION and when you consider the cheaper cost compared to Lithium technologies, their worth strong consideration.

  • Roger Franklin

    After the recent announcements from Standards Australia regarding how dangerous these batteries are, I was expecting to read that the testing centre had been burnt to the ground 4 times while testing has been underway!

    Good to see tests being completed by ITP Renewables with interesting data being produced. Just don’t tell anyone at Standards Australia that you are doing these tests in a climate-controlled room (the “Batt Lab”) at the Canberra Institute of Technology.

  • Robert Comerford

    The chemistry on some cells is stated incorrectly. The cells labelled Nickel Manganese Cobalt should be labelled Lithium Nickel Manganese Cobalt Oxide.
    For the home owner who wishes to use these in or next to their house the tests should validate the long term safety of use first and foremost.
    i.e. Just how much safer than the common fire prone LiCO2 is the addition of nickel and manganese to the mix?
    Does this stop or reduce the growth of dendrites such that an internal short will not occur during the service life of these cells?

  • I don’t see any results for the CALB cells being tested, but my own set of 400AH CALB cells has been in use for over 4.5 years now.
    I log data at 1 sec intervals (but use 5 sec interval for daily graphing), and the main change I have noted is that they will not accept as high a charging current as they used to. The absorb stage used to be a couple of hours or less, but I now need to go to 3-4 hours (or longer if I hit Vabsorb at higher charging current than typical) to get to full charge (currently at ~55.2V, with net charging current down to 5-10A).
    At a given rate of charge, I reach Vabsorb at a much lower SOC than when the battery was newer. One cell has been replaced with a Winston, due to a significantly lower capacity than the rest.

    The battery is typically discharged to about 50% SOC each night.

    Overall efficiency of the system, which I define as net amount of energy available for loads/gross charging energy out of the MPPT charge controllers, is better than 0.98. The system does get worked pretty hard at times, with around 50kWh (DC, before inverter losses) used on hot summer days, and 15-30kWh in winter.

  • Alex Hromas

    Interesting to see the high round trip efficiencies of lead acid batteries based on data about 20 years old I would have expected this to be in the high 60% range not *0%