Modeling And Parameter Identification Of Lead-acid Battery And Lithium-ion Battery

Energy storage technology can be applied to peak shaving, new energy generation grid connecting and power output, as well as enhancing grid security and stability and improving power quality, etc., which provides an effective way to solve the challenges in modern power system. Lead-acid batteries was first used in the field of electric power systems for its low price, matureness, and good reliability, etc..With the highest energy density and cycle efficiency, Lithium-ion batteries is currently the most promising high-capacity storage batteries, exhibiting a high v alue in the field of power systems. Establishment of a reasonable and accurate battery model is the assurance of the accuracy of the simulation results and the trustworthiness of the basic premise. In order to establish an accurate model of the power syste m which is suitable for research applications, reasonable structure and validity of the model is of great significance to verify identification based on measured data. As for the parameter identification and model accuracy validation, this paper conducted scientific research in aspects of battery model selection, algorithm selection, characteristics of cell structure and working principle, comparison of models parameter identification.Firstly, based on the shortcomings of traditional algorithm, parameter identification of optimization algorithms was proposed. Because the traditional optimization algorithm is sensitive to initial parameters, making identification enter the local convergence and causing the dispersion characteristics, thus,the genetic algorithm which has global search capability was chosen as the identification algorithms. Compared to the traditional algorithm, the proposed algorithm can take a greater probability of finding the optimal solution for global optimization problems. For the slow convergence and easy precocious defects in basic genetic algorithm, improved genetic algorithm was conducted in three genetic operations, such as improvement strategies, control selection parameters, and implementation. Using genetic algorithm in model identification is good in the optimization with parameters of good stability.Secondly, as for the study of lead-acid batteries, structural features, working principle and the battery characteristics wer e analyzed. For best describing cell IV characteristics equivalent, model parameter sensitivity analysis was conducted to extracts the dominant model parameters which affect the operating characteristics. Experiments proposed for the electrochemical cell m odel identification was carried out, and battery identification work based on the measured data was conducted. The actual identification of the ideal group, the average of three different parameters of the applicability and stability was analyzed. Identifi cation results show that the model can well describe the dynamic IV characteristics of the battery with good parameter stability and applicability, shows the inconsistency of battery in theoretical verificationFinally, as for the lithium-ion battery, constant steady-state current and dynamic plus sensitivity analysis was conducted, then constant power discharge and pulse discharge experimental data was used in model identification. The results show that the model can well describe the IV characteristics u nder two different conditions with good stability parameters, and the identification error is much lower than the error range of research applications requirement. Multiple factors parameter identification was conducted based on different discharge experim ental data, results showed that the charge-discharge characteristics of the model can be well simulated as battery. In the end,the impact of temperature and SOC model parameters was analyzed.