Abstract: Modeling and testing are used to characterize consequences of a first lithium-ion cell having an internal short. The vulnerability of a second lithium-ion cell being induced into thermal runaway by the energy released by the first cell undergoing an internal short is quantified. Characteristics of the packaging of Li-ion cells within a battery pack are analyzed. Combined, these analyses determine the robustness required of a cell in order to withstand a nearby cell's internal short given that the battery is maintained within the specified operational envelope by a BMS and this envelope is modified in real-time as required to meet the safety requirement. Robustness factors are: age, history of charging/discharging, as well as immediate state of charge and environment. In operation, the cell's operational history is incorporated into a model. When the model indicates cell robustness at a predetermined lower limit, operation of the cell is ceased or limited.

Abstract: A fault tolerant battery management system includes redundancy, with applications including electric vehicles. Portions of its circuitry are constituted in distinct fault domains with control, monitoring, and balancing of cells circuitry fault-effect-isolated from the circuitry associated with built-in real-time testing. Built-in tests are orchestrated in fault domains isolated from the functional circuitry being verified. These built-in tests provide test stimulus unique for each cell measurement. Cell balancing is performed in a fault tolerant manner. It takes at least two independent faults, in two mutually distinct fault domains, to negatively affect balancing capability or to interfere with a redundant circuit's ability to operate. The built-in tests allow operation without the requirement for data cross-compare between redundant measuring electronic elements. Testing and balancing functions are interlocked through encoded enabling methodologies and transmit enables on serial buses.