Design of Piecewise Affine and Linear Time-Varying Model Predictive Control Strategies for Advanced Battery Management Systems. Issue 4 (1st January 2017)

• Record Type: Journal Article
• Authors: Torchio, Marcello; Magni, L.; Braatz, Richard D.; Raimondo, D. M.
• Is Part Of: Journal of the Electrochemical Society. Volume 164:Issue 4(2017)
• Page Start: A949
• Page End: A959
• Publication Date: 2017-01-01
• Subjects: linear time varying predictive control of lithium ion batteries -- lithium-ion optimal charging batteries -- model predictive control for lithium-ion batteries -- piecewise affine predictive control of lithium ion batteries; Electrochemistry -- Periodicals; 541.3705
• DOI: 10.1149/2.0201706jes
• Languages: English
• ISSNs: 0013-4651

Abstract

Abstract : Advanced Battery Management Systems (ABMSs) are necessary for the optimal and safe operation of Li-ion batteries. This article proposes the design of ABMSs based on Model Predictive Control (MPC). In particular, we consider MPC strategies based on piecewise affine approximations (PWAs) of a first-principles electrochemical battery model known in the literature as the pseudo two-dimensional (P2D) model. The use of model approximations is necessary since the P2D model is too complex to be included in the real-time calculations required by MPC. PWAs allow to well describe the electrochemical phenomena occurring inside the battery. On the other side, the accuracy of such models increases with the number of considered partitions, which also increases the model complexity and the online computational cost of MPC. Linear time-varying (LTV) approximations, which are obtained by linearizing accurate PWAs around a nominal trajectory, are proposed as a way to further reduce online computational costs. The obtained results demonstrate the suitability of MPC based on PWARX and LTV model approximations to provide ABMSs with high performance.