‘If you don’t have the right analysis, you’ll be wasting time or money’: Powin on state of charge estimation

Traditionally, a battery management system (BMS) will use coulomb counting and the voltage-based method to estimate a battery energy storage system (BESS) SoC.

In simple terms, coulomb counting is the measurement of electric charge that flows in or out of a battery.

The voltage-based method uses the battery’s voltage to estimate how much capacity a battery has left.

The primary challenge with these methods is that the sensors used for calculating SoC are prone to error when measuring lithium iron phosphate (LFP) battery cells.

Accure Battery Intelligence and Modo Energy explain that LFPs use a flat open circuit voltage (OCV) curve, making it ‘difficult to translate voltage readings into accurate SoC values.’

They also clarify that LFP OCV curves display hysteresis effects, indicating that voltage depends not just on the SoC, but also on the direction and history of current flow.

Another factor to consider in a traditional BMS is that it does not track the battery’s ageing and is typically based on lab testing, which further increases its inaccuracy.

Powin and Tierra Climate completed an analysis on the impacts of SoC estimations to BESS operations, showing a correlation, Bennett says, particularly in energy arbitrage situations, between SoC and revenue.

Bennett talks about Powin’s research into SoC, highlighting how valuable accuracy is.

“Anywhere between 10-20%, even 30% errors of SoC have been reported historically. We think that our new algorithm, based off this six months of in the background testing, is getting us down into the low, mid-single digits of SoC estimate.”

Powin and Tierra Climate’s SoC research

A SoC algorithm that can estimate into low digits, like Powin claims, could make a massive difference in operating a BESS.

“(When SoC is inaccurate) You could have safety issues that arise if you continue to drive a system up or down beyond its safe operating parameters. As well as missed revenue opportunities and obligations.”

“Either in the moment or day to day, not providing the energy that you needed to, but also when you go to do operations and maintenance and rebalance, or fix cells that have degraded, if you don’t have the right analysis of which things to fix and when, you’ll be wasting time or money in the field,” Bennett says.

To avoid losing money, wasting time, or missing out on potential revenue, operators must choose between conservatively operating systems, which may also minimise revenue, or overestimating, which could lead to market penalties.

These kinds of losses call for an extensive amount of research in order to better operate BESS projects and avoid losses.

Bennett explains of the data collection process:

“Something that Powin did early on was to decide to collect cell-level data on our projects. Many people only collect information on the racks, which consist of generally 200 to 400 cells in series, and then various levels of aggregation within that.”

“The insights gained from cell-level data have allowed Powin to comprehend the complexities of SoC of all those batteries and how they interconnect to create what is regarded as an AC power plant, or BESS.”

Bennett continues by explaining that for these systems to operate properly, they involve the orchestration of, in some cases, millions of battery cells, fans, thermal controls and everything else that goes into a BESS.

As BESS becomes more complex, understanding how every element that comprises the system operates is critical to not just operations and maintenance, but also to SoC.

Bennett emphasises that another critical component to SoC is state of health (SoH). SoH evaluates a battery’s present capacity against its expected initial capacity to determine its remaining ability. This process allows for identifying when the cell’s performance has declined and is due for replacement.

North Carolina-headquartered System integrator FlexGen Power Systems highlighted the importance of SoH when it recently announced its battery health feature as part of its HybridOS EMS.

As part of that announcement, the company claimed that its new feature continuously evaluates the amount of energy flow during charge and discharge events to estimate the rate of degradation and computes a site-specific degradation curve.

This move from FlexGen shows how crucial extensive data and research are in better estimating SOC. Creating a robust SoH feature should, in theory, make for a more accurate SOC estimation.

Bennett says of SoH:

“You need a good SOC algorithm, but also you need to have a foundation of SoH. We think that doing that at the cell level provides a lot of unlocks as you combine these cells in different series and parallel perspectives. So, you can dig into identifying the system’s weakest or most limiting component.”

Bennett continues, “I think the combinatorial approach is really nice, where there are certain types of algorithms or approaches that will work better in certain situations, and having that data and having the flexibility to control algorithm allows you to tailor it to specific use cases or use a certain number under certain different scenarios.”

In considering data and real-life use cases, Bennett also addresses the sheer amount of data that can be obtained in these scenarios.

“The field is great because it’s real life. You just have to sort. Do that extra step to put the parameters in and search for the scenarios that meet that, and then use that, rather than sort of blindly, depending on a discharge to guide you.”

Ultimately, Bennett and Powin emphasise that though the research completed with Tierra Climate and the ability of its algorithm to estimate SOC is exciting, there are still many challenges in estimating SOC and, in turn, operating a BESS as efficiently as possible.

Bennett says that the industry needs to make progress on a standard for measuring SOC, saying:

“Because of the dynamics with LFP, you could be accurate 90% of the time and really inaccurate that other 10% of the time, but that 10% is when you’re making all your money, which has a big ramification. There’s a lot of nuance.”