The lack of a standardized approach on how State of Health (SoH) should be calculated and the fact that on-board battery management systems (BMS) alone are not sufficient in assessing battery safety are highlighting the need for battery analytics.
Knowing the State of Health (SoH) and safety status of lithium-ion batteries is fundamental for managing electric vehicle safety, maintenance, and battery warranty. Yet, keeping a battery system safe and reliable is not as easy as just operating the battery within specified thresholds. It needs far more advanced solutions to ensure seamless, safe, and profitable operation.
For battery health and safety, key performance indicators such as SoH or State of Charge (SoC), are important values to keep track of. However, determining accurate battery SoH remains a considerable challenge. There is a lack of consensus in the industry about how SoH should be calculated, meaning there is no standardized approach. Different manufacturers also calculate SoH differently. This creates complexity, particularly for fleet operators as various parties might refer to their very own SoH calculation.
On-board battery management systems (BMS) sometimes, but not always, provide an estimation of SoH. However, the accuracy of the estimations decreases over the lifetime of a battery, meaning that on-board BMS alone are not sufficient in assessing a battery’s health over its entire lifetime. These barriers to predicting battery health can be overcome by deploying battery analytics.
Lithium-ion batteries age over time, depending on how they are used. They lose capacity and increase their resistance, which, when used in vehicles, results in reduced driving performance and electric range.
The warranty provided by an electric car manufacturer usually ends when the battery reaches 70-80% of its original capacity. The time and mileage until the battery is considered ‘dead ’ must therefore be maximized, so that an electric vehicle can be operated for as many years as possible. This is necessary to increase financial returns, as well as improve the climate footprint of the battery.
At any given time, the electric vehicle owner, the car and battery manufacturer, the fleet operator, or the insurance provider may want to understand how the battery is used, and how it performs in terms of its SoH. Doing so not only allows these stakeholders to gain transparency into the health of the battery, it also enables owners and manufacturers to streamline their daily operations and protect their long-term business case.
For example, fleet operators need to know when a battery’s performance has decreased to such an extent that it will not be able to complete certain routes. SoH data is also crucial in terms of understanding when warranties can be claimed from the manufacturer.
Providing SoH data is not a standard approach, but it can be enabled by transmitting battery data from the vehicle to a cloud platform. From such a platform, the SoH can be calculated using algorithms, and can be accessed by all stakeholders if and when desired.
As well as accurately determining a battery’s SoH, ensuring safe operation should also be a top priority. If a battery malfunctions, it poses a risk to the vehicle occupants, bystanders, and potentially emergency response personnel in the case of a serious accident. Safety issues are also closely related to long-term battery performance and reliability, as such issues will often accelerate wear and degradation.
In terms of reliability, it is not uncommon for a BMS to shut down a battery at a threshold that falls below a severe safety incident. This leads to unnecessary downtime, due to the BMS’s inability to sufficiently assess the issue in question. Limiting this downtime, in addition to preventing severe safety issues, is an essential aspect of efficient battery management.
So accurately assessing battery safety and reliability is very important, but unfortunately doing so can be quite complex, as a battery breakdown or fire is usually caused by an accumulation of events over time. Eventually, these events lead to some sort of battery failure.
Thermal runaway, a situation where the battery enters a rapid self-heating state, which leads to overheating, fire, or sometimes even explosion, is a common example of such a failure. Internal short circuits are another, caused in various ways by lithium plating and dendrite growth.
Because determining the exact cause of such issues can be quite complicated, relying solely on BMS to monitor battery safety is not advised.
SoC and SoH cannot be accurately physically measured in a vehicle. Estimation models hosted by the on-board BMS estimate these by a combination of physical measurements and models, which are limited to only measuring temperature, voltage, and current.
The BMS consists of a hardware element and a software component, which keeps the battery in a ‘happy state’ at any given time during vehicle operation. It limits minimum and maximum cell voltage, current, and temperature, and is developed towards the specific battery cell chemistry, and the electrical arrangement and cooling of the modules and packs.
One of the primary objectives for the BMS is the safe and reliable operation of the battery at any given time, not the optimization over time. During operation, the BMS determines the SoC and (potentially) the SoH of the battery.
Hence, on-board BMS models for estimating SoC and SoH have limited lifetime accuracy and validation options, meaning they do not enable businesses to operate the battery in the most economical way.
The most commonly applied modeling approaches for battery SoH in the industry can bedivided into three different categories. These categories differ in the amount of data needed to parameterize the models, and in the amount of physical and chemical information involved.
While the mathematical description of electrochemical battery models is quite straightforward, the parameterization of these models is complex and requires either ab-initio calculations on an atomic basis, material-level experimental analysis, or ‘open cell ’measurements. Accessibility to these data and methodologies has improved recently, but they still require a lot of effort and computational power. Electrochemical SoH models as they are today, are not suitable for on-board BMS algorithms.
Semi-empirical models are created through a combination of cell testing and physical modeling. These models are state-of-the-art for on-board BMS estimations. They are fast and accurate enough for many applications, but they require cell testing before and during the development phase of the vehicle.
Any data-driven model gains intelligence from more data and data variability. The applicability of purely data-driven approaches for in-vehicle BMS without ‘over the air’ updates, and without the integration into a connected vehicle eco-system, is extremely limited. This is because training of the model is based solely on the one battery on board and the driving style of this vehicle.
The most promising way to assess SoH as accurately and reliably as possible is to take advantage of all three approaches. This involves exploring battery data analytics through connected vehicles, and applying physically motivated, data-driven models on one common platform. So that the information can be accessed when necessary.
For numerous reasons, relying solely on BMS does not guarantee battery safety, or allow for accurate SoH measurements.
As mentioned previously, varying definitions exist for what SoH actually means. There is no clear consensus in the industry about how it should be calculated. This means that different BMS manufacturers calculate SoH differently, and not necessarily accurately. BMS SoH algorithms also lose accuracy over time, as the BMS ages. While many risk factors are not covered in standard BMS metrics.
Overall then, BMS are generally not useful for accurately determining battery health or guaranteeing battery safety. A second layer of safety such as battery analytics is required for this. Battery analytics can provide reliable, accurate, and continuous information about the condition of batteries across an entire fleet.
Whilst on-board BMS are an important component for ensuring safe functionality of the battery, the insights they provide into battery health are limited. BMS alone are not capable of providing the data necessary to ensure battery safety.
Similarly, for fleet operators who are reliant on accurate SoH data to make important business decisions, such as planning when the battery needs to be changed, the information provided by a BMS is not sufficient.
A cloud-based battery analytics platform can provide more accurate SoH data that fleet operators can rely on. This information can be viewed at any time to provide all stakeholders with the necessary transparency.
A battery analytics platform has the added advantage of assisting fleet operators during warranty discussions, as operators and manufacturers can view battery health data in the platform, and do not need to conduct expensive testing.
.jpeg)
In the first session of the TWAICE & Camelot BESS Lifecycle Webinar Series, experts from Camelot Energy Group and TWAICE share hard-won lessons from real-world energy storage projects—helping you get BESS development right from the start.
View Webinar
%20(1).webp)
%20(1).webp)