Continuous modelling of cyclic ageing for lithium-ion batteries

Abstract

The energy industry, transportation and even the smallest consumer electronics benefit from the practical applications of rechargeable batteries. Expectations of battery performance are greatly related to capacity, power output and available lifetime. However, the lifetime is affected by gradual chemical and mechanical degradation of the internal battery structure that cannot easily be predicted prior to installation. The reduction in performance is closely related to a particular usage pattern which is unique to the user and application, and is thus difficult to predict. Reliable real-time prediction of the remaining battery life therefore remains an important research topic. In this paper we show that fading battery performance under cyclic loading can be effectively and continuously followed by introducing the concept of the damage parameter derived from mechanical durability modelling approaches. The damage parameter is calculated continuously by the novel macro-scale hysteresis damage operator model. The hysteresis model is formed by a system of constitutive spring-slider modelling elements, here bridging the complex relation between the battery load and the durability data. The spring and the slider properties are individually calibrated for lithium nickel manganese cobalt oxide (NMC) batteries, however other battery structures can also be used. The durability data is obtained experimentally under controlled steady thermal and cyclic loading (constant charge/discharge current) conditions. The approach is validated on a standardised driving pattern with a complex current history. The predicted battery life is in good agreement with observed repetitions of a simulated load block until 90% of the initial battery capacity; with 589, 590 and 698 repetitions for the combined test and simulation prediction, full simulation prediction and experiment, respectively. When compared to established equivalent circuit or analytical approaches, the proposed approach requires only a small number of cyclic durability tests with constant current and temperature. In addition, the approach supports the battery design process by allowing simulations for different usage patterns, material and durability data.