(L-1) Estimating and Managing Degradation of Li-Ion BESS Under Value-Stacked Duty Cycles in Electric Grid

Is the SoC estimated in multiple dimensions? How do you ensure its accuracy?

Yes. There are two things to consider: SoC and, of course, SoH. The main focus here is on tracking the state of health. As for SoC, we are not replacing what is already done by the battery management system (BMS) of the vehicle. Instead, we take the SoC information from the BMS and combine it with our own estimates.
Our estimates are based on a Kalman filter, which uses information from a circuit-based model, like the one I showed you, and also from coulomb counting. These two sources, when combined, typically provide a robust estimate of SoC.
Additionally, the online test I mentioned helps keep us honest by performing periodic checks—verifying the operational range under standard conditions that we run once a month. This allows us to keep the SoC and SoH estimates aligned with the true state of the battery.

For load shifting and frequency control, what are the key differences in concerns between these two modes?

Frequency control is usually very power-intensive and involves short-duration operations—typically lasting several seconds to maybe a minute or two. The power demand on the battery can be very high, whether charging or discharging, depending on whether it's frequency control up or down.  
In contrast, load shifting typically involves much longer operations, often lasting two to four hours, and the power requirements are significantly lower than those for frequency control.
So, in summary: frequency control involves higher power and shorter durations, while load shifting involves lower power, longer durations, and deeper depth of discharge. These are the key differences between the two modes, and that’s why the battery experiences different stresses under each application.

What is the frequency (e.g., quarterly, yearly, etc.) at which you calibrate your mathematical model to ensure its accuracy with real-life observations or changes?

The statistical model gives us an initial idea of how the BESS will degrade under site-specific operations. Based on that, we recalibrate the physics-based model approximately whenever we observe an additional 1% degradation—either in terms of energy throughput (kWh) or elapsed time (months).

(L-2) Battery Energy Storage Systems for Power Grids(SCiBTM)

As far as I understand, LTO has a lower energy density than graphite batteries but offers better C-rate capability. So, is the SCiB system designed specifically for frequency control mode rather than load shifting on the grid?

Our battery system is used for both frequency regulation and peak shifting. For example, the system at Nishi-Sendai is used for frequency regulation, while another system installed in Fukushima Prefecture is used for peak shifting. However, the peak shifting there is limited to about one hour.  
As you mentioned, LTO has lower energy density compared to typical Li-ion batteries. That’s why we mainly recommend using our battery system for frequency regulation. It can also be used for peak shifting, but due to cost and capacity constraints, we suggest shorter durations—one to two hours is ideal. Longer durations would make the total battery system more expensive.

What is the cooling method used?

Our battery system uses two types of cooling methods. One is a fan-based system, where a small fan is installed on the front panel of the battery—similar to those used in desktop PCs.  
The other method is air conditioning. In this case, we use airflow within the container to manage the temperature. Recently, we have shifted away from using fans because they tend to fail after 10 to 15 years, requiring replacement. To improve reliability and reduce maintenance, we now prefer using an airflow-based cooling design with air conditioning instead.

Can you explain more about thermal safety in LTO?

On page 5, at the top left corner, you can see the results of our nail penetration test. Our battery did not catch fire, even after being pierced with an iron rod.  
This is due to the materials used in our cell design. The anode material is titanium, and we use a special separator. In conventional lithium-ion batteries, when a metal object like an iron rod penetrates the cell, a short circuit occurs between the anode and cathode, causing a high current and, potentially, fire.  
However, in our battery, the separator material increases in resistance after being pierced. This limits the current flow. As a result, even after one hour, the cell retains no capacity (it is no longer functional), but the temperature remains below 100°C. This means the cell does not ignite.  
The key points are the use of titanium for the anode and a high-resistance separator. These components work together to prevent fire in case of physical damage, ensuring thermal safety.

(L-3) Stabilizing Renewable Energy Supply with Sodium-Sulfur Batteries: A Path to Carbon Neutrality

I noticed that no external cooling is required, as mentioned in the presentation. But how do you maintain the NAS battery at its desired operational temperature?

In the NAS battery, charging and discharging processes generate heat due to less-than-perfect efficiency. Typically, around 10% of the energy is lost as heat, and this heat is used to maintain the operating temperature inside the battery module.   Because the thermal insulation of the module is very good, this self-generated heat is usually enough to keep the battery at the required high temperature during regular daily operation.  
However, if the battery is only used occasionally—as in backup applications—then it does not generate enough heat on its own. In such cases, additional heater power is needed to maintain the operating temperature. That’s why we do not recommend using NAS batteries solely for backup applications.  
We recommend applications with frequent usage, where the battery generates enough internal heat to stay within the optimal temperature range.

Is the cost of NAS batteries comparable to that of lithium-ion batteries?

That’s a tough question. From a raw materials perspective, NAS batteries can be comparable to lithium-ion batteries.   However, to be honest, the situation is different when you consider the scale of production. Lithium-ion batteries—especially LFP types—are mass-produced in enormous gigafactories, particularly in China, which significantly lowers their cost.   In contrast, sodium-sulfur batteries are not yet produced at such large volumes. So, at present, NAS batteries are relatively more expensive compared to Chinese-made commodity lithium-ion batteries. That said, this may change in the future as production scales up.

(L-4) Advancements & Deployment of Flow Battery System Technology

I had a question for you. What is the voltage of the battery stack? For example, a typical lithium-ion battery stack operates around 1000 volts. How does that compare with the flow battery you're showing?

Each single cell in the flow battery has a voltage of about 1.0 to 1.4 volts. We typically connect around 100 single cells in series to form a stack. So, one stack has a voltage of approximately 100 to 150 volts between the battery terminals.

What are Sumitomo Electric's strategies for localizing electrolyte production and expanding the flow battery business in key markets without rich vanadium sources?

Most vanadium is mined in countries like China, Russia, South Africa, and Australia. On the other hand, major battery markets are in the United States and Europe. Our strategy is to source vanadium from countries like China, South Africa, or Australia, but to localize the electrolyte production in the United States and Europe. This allows us to manage supply chain risks while supporting local manufacturing in key markets.

Black start seems interesting. What are the challenges in black start, and how do you oversize the PCS output?

Since the system is isolated from the grid, it must autonomously establish voltage and frequency, requiring advanced self-control capabilities from both the battery and the PCS. In addition to simply acting as a power source, it must flexibly respond to load fluctuations. This means not only functioning as a voltage source, but also emulating grid inertia and generator behavior, which is why we employed a PCS known as a Grid Forming Inverter.