The rapid growth of battery energy storage systems (BESS) is helping to unlock the full potential of renewable energy. As battery storage is deployed at scale, increasing attention is being paid to how these assets can be operated safely and reliably throughout their lifetime. For a technology based on electrochemical processes, another important aspect comes into focus: How can risks be identified at an early stage and failures be prevented?
Dr. Cyril Marino, Failure Analysis Manager – BU Energies & Renewables at SERMA Group, discusses what failure analysis can teach us about designing safer, more resilient and longer-lasting BESS – and how testing, AI and cybersecurity can help prevent technical issues before they become costly failures. SERMA Group is an independent technology company specializing in the testing, analysis, reliability and safety assessment of battery and energy systems.
The expected lifetime of a BESS is around 20 to 30 years. To achieve this, three phases need to be considered. The first is the design phase, where risks and potential problems need to be addressed from the outset. The second is operation, where the system needs to be continuously monitored. And the third is maintenance: if a component fails, it needs to be taken out of service and replaced in time. The key is to prevent failures from turning into financial losses.
At SERMA, we do failure analysis and it is our job to learn from mistakes.
We see that thermal management is still the hot topic and one of the high risks.
Battery cells are highly sensitive to temperature, and parts of a cell that operate at a higher temperature will age faster. Battery charging also generates heat, and the resulting temperature differences cannot always be controlled by the BMS (battery management system).
When looking at the whole system, complementary components such as cooling must also be taken into account. They require electricity, and when this electricity comes from the battery, it creates a parasitic load. In some installations, for example, solar panels provide power for the cooling system. When there is no solar generation at night, the battery must supply the cooling itself. This can be risky if there is not enough energy available to keep it running.
A new and very critical factor for BESS is cybersecurity. Modern BESS are increasingly connected to remote monitoring, energy management and supervisory control systems. This connectivity creates additional cybersecurity attack surfaces. Phishing and compromised user credentials can provide an entry point into corporate or operational networks connected to the BESS. There are small risks like data theft. But the big risk is that someone gains access to the network and modifies control parameters, disrupts operation or causes the system to shut down.
We have seen with power systems that one problem can create a domino effect: if something goes wrong at one point, it can affect other parts of the system as well.
Furthermore, the location of data storage and control can represent a geopolitical risk. For example, with some inverters from China, the database is controlled by the manufacturer there and the data may be handled through servers located abroad.
There are different ways to address risks when designing the BESS. With thermal management, for example, different cooling systems can deliver cooling directly to the point in the battery where it is needed, while additional thermal sensors can improve the warning system and help identify potential issues earlier. Where possible, it is also better to connect the cooling system to a separate power source, in order to avoid parasitic load.
Battery technology makes a difference, too. Technologies such as NMC (nickel manganese cobalt) are able to create oxygen during thermal runaway of batteries, whereas it is not the case with LFP (lithium iron phosphate) chemistry. Thermal runaway is a self-reinforcing overheating process within a battery cell. With NMC batteries, temperatures could exceed 1000 °C during the incident, while temperatures in LFP batteries can remain below 600 °C, as a consequence. The choice of chemistry should therefore be based on the specific requirements and conditions of the application.
In case something happens, the appropriate fire extinguishing system is another key consideration. The right system needs to be selected carefully rather than simply choosing one because it meets the standard. It is important to understand that batteries might produce oxygen when burning. A system that is simply designed to reduce the oxygen level might not work if a battery catches fire.
Cybersecurity, on the other hand, starts with the right training of people. They must be enabled to understand and handle these risks. To protect the BMS, one effective measure is network segmentation. If a failure or cyberattack occurs in one part of the system, segmentation can prevent the problem from spreading to the entire installation or wider network. Furthermore, data storage and control should be managed in secure environments to reduce potential geopolitical risks.
Europe needs its own data centers for this kind of critical infrastructure.
Testing and validation help identify the limits and potential weaknesses of a BESS before problems occur. It is not sufficient to rely solely on generic standards. Each system has its own characteristics and should be tested under different operating and environmental conditions, for example at high temperatures such as 40 °C or with higher and lower currents. This helps identify where the limits are and how the system can be made more robust.
Battery health monitoring is increasingly benefiting from artificial intelligence. Traditionally, mathematical models have been developed based on experience, monitoring, laboratory testing, experiments and feedback from operation. With AI, we can now combine different kinds of mathematical models from different fields and improve them using operational data.
For example, we can monitor temperature, voltage and current, as well as the internal resistance of the battery. AI can use this data to identify potential risks at an early stage. This is a big revolution in the field. The monitoring system can become very close to predicting what might happen, which gives us the possibility to extend the lifetime of the system.
Absolutely. Extending battery lifetime makes sense from both the economic and environmental point of view. The battery is going to be recycled in the end, but its lifecycle should be prolonged as much as possible to get the most out of the battery. This can lower costs and have environmental benefits.
For example, car batteries can still deliver energy after they are no longer used in the car. The question is how much lifetime is still available. The final usage needs to be assessed carefully due to potential risks. In some cases, a certificate is needed, which can be costly. It is still questionable whether this is economically valuable in every case.
In addition, the new battery regulation states that new batteries accepted in the EU need to have a certain percentage of recycled content. Recycling will therefore become an integral part of the battery's life cycle.
Climate risks will play a key role. BESS need to be resistant to wind, air and heat and prepared for extreme climatic conditions. In combination with PV systems, batteries need to cope with volatility and adapt to changes in generation. They also need to feed electricity into the grid when it is needed.
The main point is that battery usage must be optimized to get the most out of the system while maintaining its lifetime and safety. Artificial intelligence will contribute significantly to increasing efficiency, optimizing battery life and creating additional safety barriers. In the end, this is also about the profitability of the system.
