NiMH Battery Characterization and State-of-Charge Estimation for HEV Applications

Abstract

One of the most important challenges for effective energy management in Hybrid Electric Vehicles (HEVs) applications is the capability of correctly estimate the batteries State of Charge (SoC). This task is particularly arduous to accomplish in real-time, due to the complexity and nonlinearities, as well as the inevitable presence of on-board measurements errors. The objective of this paper is to develop a NiMH battery model which accounts for the relevant dynamics related to HEV applications. A detailed series of experiments allowed for system identification to be performed in order to design an equivalent electrical circuit model of the battery, where capacitors take care of voltage relaxations, while resistors model all forms of energy losses. Experimental results are then compared with model prediction, resulting in variances limited to less than 3.5%. Finally, the validated model was inverted to serve as a main component of a proposed SoC estimator, whose performance is of key relevance for effectiveness of HEVs. Results during actual vehicle operations are shown delineating a successful application.

Electric vehicle battery model identification and state of charge estimation in real world driving cycles

“…There is a general agreement of the results of this study with previous studies on NiMH battery parameterisation [5]. In addition, the identified model is validated against the experimental data.…”

Low-cost programmable battery dischargers and application in battery model identification

“…As shown above, the measurements follow a stable pattern. In addition to the general agreement of the measurements with previous studies on NiMH battery parameterisation [20], the results do appear useful, providing a consistent way to estimate model parameters.…”

Model-based calibration for battery characterization in HEV applications

“…In [1], a model-based SoC estimation is presented, which blends a simple current integration with an estimate of the SoC based on a model inversion. The model used in the inversion is an interesting structure that combines a simple electrical circuit with a set of dynamic open circuit voltage maps.…”

“…The study in [1] suggests that the hysteresis behavior of the OCV can be successfully captured combining an equivalent electrical circuit with different OCV maps while providing a physically meaningful smooth transition from one map to the other, as well as suitable relaxation. Because optimization is used, we do not need to construct accurate OCV maps.…”

“…Taking into consideration all of the above considerations, a first order equivalent electrical circuit model, adapted from [1], is chosen as shown in Figure 1. The diodes shown are ideal diodes and are only used to reflect the fact that the resistances in the RC circuits are different during charging and discharging.…”