modeling, real-time degradation identification, and remediation of lead-acid battries

Open Access
Shi, Ying
Graduate Program:
Mechanical Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
October 07, 2013
Committee Members:
  • Christopher Rahn, Dissertation Advisor
  • Christopher Rahn, Committee Chair
  • Chao Yang Wang, Committee Member
  • Hosam Kadry Fathy, Committee Member
  • Jeffrey Scott Mayer, Committee Member
  • modeling
  • battery system management
  • degradation diagnosis
  • desulfation
Valve Regulated Lead-Acid (VRLA) batteries are cheap and mature technology, favorable candidates for micro-hybrid vehicles and stationary applications. Large-scale battery packs, instead of individual cells, are implemented in those applications; therefore sophisticated battery management system (BMS) becomes necessary and crucial to ensure the longevity and efficient utilization of battery packs. Such a BMS must have several key elements: a simple but accurate system model that captures battery performance and aging processes, sensor measurements that gather information to help the controller monitor battery health and identify major aging mechanisms, and an advanced controller that estimates state of charge (SOC) and state of health (SOH) of cells and optimizes their usage accordingly. This research first reviews six modeling techniques that are suitable for developing electrochemistry-based system models of batteries. Fundamental battery models, consisting of nonlinear coupled partial differential equations, are often difficult to discretize and reduce in order so that they can be used by systems engineers for design, estimation, prediction, and management. In this work, six methods are used to discretize a benchmark electrolyte diffusion problem and their time and frequency response accuracy is determined as a function of discretization order. The Analytical Method (AM), Integral Method Approximation (IMA), Pade Approximation Method (PAM), Finite Element Method (FEM), Finite Difference Method (FDM) and Ritz Method (RM) are formulated for the benchmark problem and convergence speed and accuracy calculated. The PAM is the most efficient, producing 99.5\% accurate results with only a 3rd order approximation. IMA, Ritz, AM, FEM, and FDM required 4, 6, 9, 14, and 27th order approximations, respectively, to achieve the same error. If both modeling complexity and efficiency are considered, Ritz method is the best candidate. Secondly, this research presents a nondestructive experiment method to perform real-time aging diagnosis of lead-acid batteries. VTLA batteries can degrade due to a variety of mechanisms, including corrosion, hard sulfation, water loss, shedding, and active mass degradation. VRLA batteries are designed to minimize these effects as much as possible but the operating environment, cell-to-cell and battery-to-battery manufacturing variations, and use can cause different degradation mechanisms to dominate capacity loss and/or impedance rise. With accurate State of Health monitoring, cell usage can be adjusted by the battery management system (BMS) to optimize the performance and life of the energy storage system. The BMS must be able to determine in real time the predominant degradation mechanism for each cell and adjust use accordingly. In this work, new and dead VRLA batteries are tested with constant, sinusoidal, and pulse charge/discharge current inputs while measuring the cell voltage and pressure to determine the cause of death of the cells. As expected, the new cells have fairly uniform performance with limited signs of degradation. The cells in the dead battery, however, have widely ranging performance, especially at the end of discharge and charge. Analysis of the charge/discharge data indicate that three cells died of water loss and a fourth cell died of sulfation. The remaining two cells were fairly healthy but will accompany their dead companions to the recycling center nonetheless. While the full charge/discharge data provided useful forensic pathology data, EIS and pulse charge/discharge data varied with aging mechanisms and only provided supplementary pathology information. Following the real-time diagnosis work, a charging control scheme is proposed that removes hard sulfation in lead-acid cells without introducing excessive gassing. In a battery string, the cell with the lowest capacity dominants that of the entire string. If that cell's capacity can be recovered, the capacity of the whole string will increase. However, not all aging mechanisms in lead-acid batteries are reversible but hard sulfation is. Often, removal of one degradation mechanism might worsen another. In this study, it appears that one cell of a 6-cell string died from sulfation and another three from dehydration. The battery capacity is mainly dictated by the sulfated cell. A desulfation charging control scheme with pressure feedback is designed to break up hard sulfate and recover capacity while minimizing water loss by using low current charging. The capacity of the cell is recovered by 41\% with minimal water loss, demonstrating the effectiveness of the desulfation charge controller. The experiments reveal the great potential of charge strategies with pressure control. To make this health-cautious charge more cost-effective and easy to implement, a nonlinear system model is developed, aiming to eliminate the pressure transducers by covering gassing side reactions in the model. The system model is fifth-order with parameters and states that are based on the electrochemical processes and battery properties. It preserves the majority of the underlying complex mathematical model but enjoys the beauty of state space form, which eases future controllers and estimators design. The model is validated with testing data and shows good match in battery voltage and pressure responses. It also returns the internal states such as acid concentration, solid-phase potentials, and transfer current densities. Those states can be used for control in the future. Powered by the nonlinear system model a health-cautious charge strategy, constant-current constant-overpotential (CC-Ceta), is presented. In a lead-acid cell, the overpotentials directly control the oxygen and hydrogen generation. Compare to conventional constant-current constant-voltage charge protocols, CC-C$\eta$ charge protocols produce less gas and increase the charge efficiency. Controlling eta on the positive electrode directly suppresses the oxygen generation and also affects eta on the negative electrode, slowing down hydrogen production, vice versa. It has also been shown that a SOC estimator based on the internal acid concentration level has better performance than one that is based on current counting in the overcharge stage, where SOC curve is high nonlinear and part of the input current goes to side reactions.