Improve in-situ life prediction and classification performance by capturing both the present state and evolution rate of battery aging
This study develops a methodology by capturing both the battery aging state and degradation rate for improved life prediction performance. The aging state is indicated by six physical features of an equivalent circuit model that are extracted from the voltage relaxation data. And the degradation rate is captured by two features extracted from the differences between the voltage relaxation curves within a moving window (for life prediction), or the differences between the capacity vs. voltage curves at different cycles (for life classification). Two machine learning models, which are constructed based on Gaussian Processes, are used to describe the relationships between these physical features and battery lifetimes for the life prediction and classification, respectively. The methodology is validated with the aging data of 74 battery cells of three different types. Experimental results show that based on only 3-12 minutes' sampling data, the method with novel features predicts accurate battery lifetimes, with the prediction accuracy improved by up to 67.09% compared with the benchmark method. And the batteries are classified into three groups (long, medium, and short) with an overall accuracy larger than 90% based on only two adjacent cycles' information, enabling the highly efficient regrouping of retired batteries.
Code (0)
등록된 구현이 없습니다.
Tasks
Gaussian ProcessesPredictionSimilar Papers 제목 키워드 기반
ELC: Evidential Lifelong Classifier for Uncertainty Aware Radar Pulse Classification
Reliable radar pulse classification is essential in Electromagnetic Warfare for situational awareness and decision support. Deep Neural Networks have shown strong performance in radar pulse and RF emitter recognition; ho…
Continual LearningA Multi-Level Hierarchical Framework for the Classification of Weather Conditions and Hazard Prediction
This paper presents a multilevel hierarchical framework for the classification of weather conditions and hazard prediction. In recent years, the importance of data has grown significantly, with various types like text, n…
Self-Driving CarsWeather ForecastingEmbedded System to Detect, Track and Classify Plankton Using a Lensless Video Microscope
Plankton provide the foundation for life on earth. To advance our understanding of the marine ecosystem, for scientific, commercial and survival purposes, more in situ continuous monitoring and analysis of plankton is re…
Raspberry Pi 3Few-Shot Lifelong Learning
Many real-world classification problems often have classes with very few labeled training samples. Moreover, all possible classes may not be initially available for training, and may be given incrementally. Deep learning…
Continual LearningLifelong learningMulti-Object Tracking Consistently Improves Wildlife Inference
Camera traps have become a common tool for wildlife monitoring efforts in ecological research and biodiversity conservation. Wildlife classification models have benefited from the increase in wildlife visual data. These …
Multi-Object Tracking