BESS Battery Cell Fault Detection With Adaptive Outlier Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery energy storage systems (BESS) face challenges in accurately detecting and classifying abnormal battery cells due to nonlinear time-varying dynamics, leading to potential catastrophic failures from undetected faults, with current methods often resulting in false alarms or late detections.
Innovation Solution
A system and method using a battery management system (BMS) to collect data from BESS batteries, which employs a prognostics and fault detection model trained by a data classification neural network, incorporating adjacency weighted, temporal, and spectral distance informed classifiers to identify outlier battery cells and classify potential faults.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If rule-based methods with fixed thresholds are used to detect abnormal battery cells, then false alarms are avoided by conservative threshold setting, but detection sensitivity decreases and abnormal situations are detected late
Solution Approach 1:
The patent transforms fixed threshold parameters into dynamic thresholds that adapt to battery operating conditions. The system learns normal voltage and temperature ranges through machine learning models, allowing thresholds to change based on state of charge, temperature, and operational context, thereby maintaining high detection sensitivity while controlling false alarms
Solution Approach 2:
The patent replaces simple rule-based mechanical threshold checking with intelligent systems using machine learning algorithms. The system uses historical data training to create adaptive detection models that can identify abnormal patterns beyond fixed thresholds, improving detection precision without sacrificing reliability
2Productivity
If automated methods are employed to identify abnormal voltages or temperatures, then detection speed improves, but false alarms increase due to marginal deviations from normal values
Solution Approach 1:
The patent moves from single-parameter detection to multi-dimensional analysis by simultaneously monitoring voltage, temperature, current, and their rates of change. The system analyzes patterns across multiple dimensions and time, using contextual information to distinguish true abnormalities from normal variations, thereby reducing false alarms while maintaining fast detection
Solution Approach 2:
The system implements feedback mechanisms where detection results and operational data are continuously fed back to refine the machine learning models. This allows the system to learn from past detections, adjust sensitivity parameters, and improve accuracy over time, reducing false alarms while maintaining high detection speed
3Reliability
If multiple BESS containers are deployed in a microgrid, then power supply reliability improves, but monitoring and analysis complexity increases significantly
Solution Approach 1:
The patent creates a universal monitoring platform that can manage multiple BESS containers through a single integrated system. The machine learning models are designed to handle data from multiple containers simultaneously, applying the same detection algorithms across all units, thereby simplifying monitoring complexity while maintaining reliability across the entire microgrid deployment
Data Source
AI summary
A system and method for detecting and classifying outlier battery cells operating abnormally in a storage battery of a battery energy storage system (BESS). A controller controls the operation of the BESS, and a battery management system (BMS) collects battery operational data from the storage battery and stores the battery data in a data repository. A prognostic agent coupled to the battery data repository uses the stored battery data to train a prognostics and fault detection model that is loaded in the controller and used to detect at least one outlier battery cell. Detected outlier battery cell and their operational data are classified using a data classification neural network to one of a plurality of fault types.