Research On SOC Estimation Based On Fusion Algorithm Of IPSO-EKF

Accurate SOC estimation of lithium-ion battery can ensure the health of power battery and avoid life degradation and safety problems caused by overcharge and overdischarge.Considering that the underlying model of KF（Kalman Filter）is relatively rough and the noise of system and observation is out of consideration,a novel fusion algorithm combining EKF（Extended Kalman Filter）and IPSO（Improved Particle Swarm Optimization）,namely IPSO-EKF,is proposed in this work to optimize noise covariance matrix,with the improvement of SOC accuracy and the expansion of applicability under distinguished profiles.The principal components and academic researches of this paper are threefold:（1）Loads of experiments are designed and carried out to learn about the characteristic parameters of the battery according to the external characteristics of the Li Fe PO₄ cell,with the static capacity test to calibrate the maximum capacity and the open circuit voltage test to calibrate the mapping relationship between SOC and OCV,as well as the hybrid power pulse characteristic test to identify the model parameters and diverse dynamic profiles to evaluate the accuracy and applicability of the algorithm,which provides the experimental basis for the subsequent verification of the proposed algorithm.（2）To address the problem that the underlying circuit is relatively crude and humble,a novel third-order equivalent circuit model is employed to describe the volatile battery dynamics caused by electrochemical polarization reaction and concentration difference polarization reaction,and simulate the external characteristics of power battery.Moreover,the off-line method is selected to distinguish the model parameters,with the smoothing spline function to map the corresponding correlation between SOC and each equivalent impedance,and the accuracy of the fitting curves is evaluated through mathematical characteristics later afterwards.Furthermore,the accuracy of model parameters is verified by DST,UDDS and NEDC.（3）The principle of EKF algorithm is demonstrated,pointing out that the deficiency of noise covariance matrix optimization,and later,the influence of random value of noise covariance matrix on the estimation of SOC is discussed under static and dynamic profiles.Subsequently,the principle of PSO is described,pointing out that the characteristics of population optimization are suitable for noise optimization,and a SOC estimator is proposed based on IPSO-EKF fusion algorithm under the premise of the improvement of PSO,with the noise covariance matrix of system and observation optimized during the EKF iteration incorporated IPSO algorithm under diverse profiles,so as to improve the accuracy of SOC estimation.After that,the simulation curves and error curves between EKF and IPSO-EKF are compared and analyzed under static and dynamic profiles,respectively.In addition,the reliability of the proposed algorithm is evaluated by RMSE（Root Mean Square Error）and MAPE（Mean Absolute Percentage Error）.