| Energy storage
predictive battery analytics that combine physics- based modelling, portfolio-wide intelligence, and long-term historical data analysis to produce accurate SOC estimates that the BMS alone cannot
deliver. Three core capabilities make this possible: ● cloud computational power; ● historical data; and ● comparisons across the whole asset portfolio.
● Cloud computational power. Cloud-based analytics draw on significantly more computing power than a traditional BMS. This enables advanced physics-based models to be deployed that can resolve the flat open-circuit voltage curve that plagues LFP systems and is one of the main drivers of SOC inaccuracy. The result is precise mapping of voltage to SOC, especially useful in the mid-range SOC where BMS algorithms routinely lose accuracy. ACCURE’s SOC algorithm is built on a foundation of Coulomb counting combined with voltage-based recalibrations, the industry standard for BMS SOC estimation. It implements significantly more recalibrations using an advanced, field-data-driven electrical model. For LFP cells, this is enhanced with a hysteresis model that provides many more orientation points beyond the typical full charge and discharge references used in standard BMS systems. On top of this, machine learning models, including neural networks trained on large-scale field data, are used to validate and refine the physics-based outputs, adding an additional layer of accuracy and confidence.
● Processing historical data. The BMS is short-sighted by default. It operates iteratively, processing a new set of measurements at each time step, and then immediately calculating the next step. While the BMS does operate in real-time, it has limited ability to process larger coherent datasets and track long-term trends. This means it lacks context to draw accurate conclusions.
Cloud-based analytics, on the other hand, have access to all the historical data from a battery, making it easy to monitor long-term trends, and detect and compensate for issues like SOC drift from Ah-counting (as shown in Figure 2). Cloud-based predictive battery analytics can detect and quantify this trend by estimating the offset current causing the SOC estimate to drift away from reality. Once this trend is identified, it can be compensated for, resulting in a corrected SOC curve with Ah-counting (light green trace, Figure 2). Instead of reaching a 10% SOC error after 14 days, this method ensures that the cloud-computed SOC remains accurate for weeks and even months without needing frequent recalibration.
Beyond drift compensation, historical data also enables accurate internal resistance and state- of-health (SOH) calculations. These are both long-term KPIs that directly determine how much energy can be extracted from a battery and are otherwise difficult to derive reliably from the BMS’s step-by-step view alone. These capabilities enable cloud-based predictive battery analytics to achieve accurate SOC estimates within 2% of the actual value.
Time (days)
Figure 2. Illustration of SOC drift that can be compensated for by cloud-based analytics. Source: ACCURE.
● Comparisons across the whole asset portfolio. A primary challenge for a BMS is that it only monitors the battery it is connected to. Even an advanced BMS capable of learning from past operations remains limited to data from its own single battery. But what happens when that battery encounters conditions it hasn’t experienced before, like extreme temperatures or a safety-critical state? In this case, it has no data from which to draw informed conclusions on SOC.
In contrast, cloud-based predictive battery analytics can tap into data from across an entire portfolio of battery assets to offer unprecedented accuracy in SOC estimation. Predictive analytics can transfer insights gained from one battery to the entire network, enabling comparative analysis that quickly identifies outliers. By benchmarking SOC errors across all deployed batteries, cloud-based analytics can zero in on the exact conditions that are causing the BMS to inaccurately estimate SOC, identify these patterns and help mitigate them to determine an accurate SOC.
Using cloud-based SOC estimation to boost revenue
Inaccurate SOC estimation can lead to significant financial consequences, especially for a BESS engaged in energy markets and grid services. To capitalise on cloud-based insights, they must be actively integrated into the operational strategy.
One approach is to manually adjust the trading algorithm using cloud-provided offsets to correct SOC discrepancies. While this can improve trading accuracy and reduce financial risk, it is only a partial solution. The true value of analytics is only realised when the trading
algorithm is fully integrated with the cloud platform, allowing continuous, near-real-time updates of the SOC. This ensures that trading decisions are always based on the most accurate and up-to-date data, delivering these
commercial benefits: ● more precise dispatch and better spread capture;
● minimised risk of penalties for non-compliance with contractual obligations;
● enhanced revenue streams as more of the energy stored in the BESS can be sold; and
● optimised asset utilisation, which reduces the conservative buffers that eat silently into capacity and revenue.
At its core, integrating advanced analytics with the trading algorithm provides a dynamic and responsive approach to SOC management, turning a potential liability into a competitive advantage in the energy market.
In summary: cloud-based predictive battery analytics turn a potential liability into a competitive advantage Inaccurate SOC is not an inevitable drawback of operating LFP-based BESS. Cloud-based predictive analytics, backed by physics-based modelling and portfolio-wide intelligence, can now bring SOC estimation accuracy to within 2% of actual value, and the commercial case for making the change is clear. Operators that integrate SOC Correction into their trading workflows dispatch with greater precision, reduce deviation risk, and recover revenue that BMS inaccuracy has been quietly eroding. The assets are already built. The question is whether they’re earning what they were built to earn.
https://www.accure.net
www.modernpowersystems.com | May/June 2026 | 33
State-of-charge (%)
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45
