Estimating Imbalance is Critical - How Do We Know Our Signals are Accurate?

One of the fundamental limitations of conventional battery energy storage system (BESS) control systems is unknown imbalance levels—and the consequences of this limitation are serious. Left unaddressed, imbalance can dramatically compromise your BESS assets’ availability, performance, and profitability. 

The reason BESS controls usually do not report the imbalance levels of an asset, despite its importance, is that imbalance signals are difficult to estimate and even more difficult to validate.

Zitara's advanced imbalance estimation algorithms are precise, robust, and provide real time insight to operators, allowing assets to reach their maximum performance with minimum balancing downtime.

Since imbalance inside an asset is invisible and cannot be directly measured, Zitara pays particular attention to how we validate the performance of our algorithms. Any vendor's control signals are only valuable if they are reliable—here, we explain how we ensure our algorithms are highly accurate and meet real-world needs even in challenging environments.

‍

The challenge in inferring rack-level imbalance

To determine the imbalance level of a rack in your BESS, you typically measure the difference in state of charge (SoC) between the cells with the highest and lowest voltages in the rack. But determining accurate SoC estimates is where the real challenge lies. 

While using SoC lookup tables that correlate cell voltage to SoC is the most common approach to inferring imbalance levels, it’s not necessarily the most reliable for real-world BESS operations. 

SoC lookup tables can only deliver high-confidence estimates under specific conditions: when the battery has been at rest for hours at a low SoC. In real-world BESS operations, these moments (indicated as yellow markers in the below plot) are quite rare as the system is typically either charging, discharging, or otherwise operating. 

Even worse, these points can only be identified in the past, not in real time, making it infeasible to use them in real time controls. Without performing a careful retrospective analysis, other less reliable measurements (indicated in gray) "throw off" a control system's signals.

As a result, using SoC lookup tables to make estimates is often inconsistent, unreliable, and misleading. 

Inconsistency is only one reason why SoC lookup tables fall short. Discover the 5 reasons why SoC lookup tables can’t support BESS.

‍

‍

Zitara uses past and future data to validate our algorithms

How we validate Zitara Balance

To ensure the accuracy of Zitara Balance, we assess our output with inferred imbalance levels derived using the SoC lookup table approach—but only after verifying that the data points are accurate and measured under ideal conditions by looking forward and backwards in time. 

While rare, these conditions do provide a high-fidelity baseline and can thus give us a confident inference of the true imbalance levels. A retrospective analysis, which uses the future behavior of the asset to confirm that the baseline is accurate in the past, gives us the true imbalance of the asset over time.

Once this ground truth baseline has been identified, we rigorously test our algorithm using a data-based backtest, where the algorithm only sees data in simulated real time, and must make an accurate determination without the benefit of foresight.

By leveraging verified imbalance data over large-scale backtest data sets, we can confidently validate the accuracy of Zitara Balance and its ability to deliver reliable, actionable insights you can use to enable smarter balancing strategies across your entire fleet.

‍

How Zitara Balance works

Zitara Balance uses the cell-level telemetry your assets already track (e.g., voltage, current, and temperature) and transforms it into real-time rack- or string-level  balance/imbalance signals using an accurate model of the true behavior of the cells driving the imbalance. It also provides two key metrics: 

• State of Balance: a snapshot of how balanced each rack is
• Unavailable energy: how much energy operators can recover from each rack, block, and project with targeted balancing activities

In turn, these signals enable rack-level prioritized balancing, which minimizes downtime and maximizes system availability. 

‍

A better way to measure and manage imbalance

Imbalance is a persistent challenge in BESS operations that can quietly but significantly undermine asset availability, performance, and ultimately, profitability. SoC lookup tables remain the most common method for inferring imbalance levels, but their finicky results make them unfit for the demands of real-world, in-the-field BESS operations.

Zitara Balance, on the other hand, seamlessly integrates with your existing control systems, consistently providing accurate, real-time visibility into rack-level imbalance to enable smarter balancing strategies that increase asset availability, improve performance, and maximize revenue generation.

Learn more about Zitara Balance and how it can help you get more from your BESS assets in our white paper: Maximizing Asset Availability with Zitara for BESS.