The lithium-ion battery (LIBs) has become one of the optimal choices for energy storage systems (ESS) in many fields, such as electric vehicles (EVs) and microgrids.
However, an ESS which only depends on batteries has several drawbacks: (1) In order to meet the peak power demand, the power density of the batteries needs to be high enough. Although batteries with high power densities have been developed, the cost is much higher than the common ones. (2) The power improvement of the ESS needs large-scale batteries and a high discharge current, thus, batteries are more likely to suffer from thermal failure, over-discharge, and other failures.
On the one hand, the temperature should be strictly controlled in order to ensure safety under high power-load conditions. On the other hand, long operating times with high-rate charge/discharge current will cause accelerated aging and capacity loss. (3) In some applications (e.g. EVs, HEVs, etc.), batteries usually suffer from frequent charge and discharge, which has an adverse effect on battery life. In order to enhance the operation of ESS, supercapacitors are designed to integrate with batteries as buffers which make this hybrid energy storage system (HESS) more robust in handling surge currents.
In the LIB/supercapacitor HESS, the supercapacitor provides a fast and effective energy output because of its high power density and high efficiency. However, the supercapacitors have a low energy density, as they only store energy by surface adsorption reactions of charged species on an electrode material. Fortunately, the LIB has complementary characteristics with high energy density but relatively low power density. This hybridization is considered as a better use than individual use of LIBs or supercapacitors, and it provides a large advantage over the two devices.
To guarantee safe operation of the HESS, an energy management system is necessary in order to provide intelligent functions, including accurate states estimation, thermal management, fault diagnosis, etc. To ensure the LIBs and supercapacitors work in a suitable area and prevent them from over-charging and over-discharging, accurate and real-time estimation of power capability and maximum charge and discharge energy is crucial and necessary in the energy management system.
In this work, a model based multi-timescale power capability and maximum charge/discharge energy prediction approach is presented (see fig.1). An explicit analysis of power capability and maximum charge/discharge energy prediction with multiple constraints such as current, voltage, and state-of-charge (SOC) is elaborated. In order to overcome estimation errors caused by sensor noises, the extended Kalman filter is employed for power capability and energy prediction. The charge and discharge power capability and the maximum charge and discharge energy are quantitatively assessed under different dynamic characterization schedules.
In the first part of the analysis, the model framework of the LIB and supercapacitor hybrid system is developed, which consists of three parts: the LIB, the supercapacitor, and the DC/DC converter. Then, the mathematical tool for model parameter identification is introduced. The data-driven algorithm of the recursive least-squares method with forgetting factor is explained in detail.
In the second part of the analysis, the multi-timescale power capability prediction considering multiple constraints has been discussed for both the LIBs and supercapacitors. In the power capability estimation, one of the emphases is the state estimation of SOC. Since the Kalman filter has been applied extensively to the field of system state estimation, which provides an efficient recursive way for parameter prediction, the Kalman filter is employed for power prediction. Finally, the maximum charge/discharge energy for the LIBs and supercapacitors are calculated based on the power capability prediction results.
In the last part of the analysis, to validate the precision of the developed models and the robustness of the proposed power and energy prediction method, the validation experiments under different dynamic characterization schedules are carried out and the results are analyzed. The results indicate that accurate voltage predicted values can be obtained by the presented models and parameter identification algorithm. The maximum charge and discharge energy with different time scales for both the LIBs and supercapacitor are compared. From the results, we can conclude that the prediction with small time scales has smaller absolute values of maximum charge and discharge energy, which means that long timescale maximum charge and discharge energy estimation are not suitable for longtime applications or else the battery will be over-charged or over-discharged.
These findings are described in the article entitled Multi-timescale Power and Energy Assessment of Lithium-ion Battery and Supercapacitor Hybrid System using Extended Kalman Filter, recently published in the Journal of Power Sources. This work was conducted by Yujie Wang, Xu Zhang, Chang Liu, Rui Pan, and Zonghai Chen from the University of Science and Technology of China.