Energy Storage System (ESS) Anomaly Detection: 12 Hours in Advance

  • Energy
  • Predictive Maintenance

Challenge

Energy Storage Systems (ESS), which hold substantial energy, are prone to generating excessive heat during repeated charging and discharging cycles. This can lead to large-scale fires if cells accumulate fatigue or have defects. Therefore, there is a critical need for a system that can detect such anomalies early to ensure safety and prevent accidents.

Approach

The project involved a detailed analysis of the thermal runaway process in Energy Storage Systems (ESS). This was accomplished by utilizing operational data from ESS under normal conditions, combined with specialized battery anomaly test data obtained from a dedicated battery research laboratory. Leveraging this comprehensive dataset, a predictive model was developed. This model, crafted through the use of semi-supervised anomaly detection techniques, was specifically tailored to anticipate failure scenarios as outlined by customer specifications.

Value Delivered

The newly established safety metrics enabled proactive responses, allowing the ESS to initiate preventive measures 30 minutes prior to the potential onset of thermal runaway. Furthermore, the introduction of advanced rack drop precursor detection models proved pivotal in early anomaly detection. These models effectively predicted and monitored for early signs of potential fires, offering a substantial lead time of up to 12 hours for preemptive action, thereby markedly improving the overall safety and reliability of the ESS operations.

Want to learn more about this use case?
Get in touch with our industrial AI experts.


Talk to an AI Expert