Title :
An integrated data mining framework for analysis and prediction of battery characteristics
Author :
Momtazpour, Marjan ; Sharma, Ritu ; Ramakrishnan, N.
Author_Institution :
Dept. of Comput. Sci., Virginia Tech., Blacksburg, VA, USA
Abstract :
Batteries play an important role in modern sustainable energy systems. However, batteries are expensive and have a limited life time. Having a deep understanding of how batteries operate in working situations is crucial to designing advanced control mechanisms. Battery performance and life time is highly dependent on how it is used and also on environmental working conditions. While batteries have been extensively studied through model-based approaches, there is no previous work about modeUng behavior based on data analytic methods. In this paper, we propose an integrated data-driven framework to study the behavior of battery systems in a grid, based on data mining techniques. The proposed method provides a high level characterization of battery behavior and online parameter estimation using supervised and unsupervised learning methods. This work can be used in intelligent control systems and would help administrators to know what is happening inside a battery.
Keywords :
battery storage plants; data analysis; data mining; intelligent control; sustainable development; unsupervised learning; battery life time; battery systems; data analytic methods; data mining; intelligent control systems; sustainable energy systems; unsupervised learning; Asia; Batteries; Clustering algorithms; Degradation; Discharges (electric); System-on-chip; US Department of Defense;
Conference_Titel :
Innovative Smart Grid Technologies - Asia (ISGT Asia), 2014 IEEE
Conference_Location :
Kuala Lumpur
DOI :
10.1109/ISGT-Asia.2014.6873891