Parameter identification method for the variable order fractional-order equivalent model of lithium-ion battery

Abstract

The equivalent model structure, fractional order-value, and basic circuit parameters in a fractional-order equivalent circuit model (FECM) of a battery jointly determine the description ability of the FECM. For a certain model structure, the fractional order-value determines the basic form of frequency domain impedance characteristics of the model. However, in practical applications, the determination of the fractional order-value lacks clear guidelines. In this paper, the orders of FECM are taken as parameters to be identified, and the coevolutionary particle swarm optimization algorithm is used to synchronously identify the parameters and order-values of equivalent model from a complete set of load profile test data. Under the same algorithm conditions, the state description capabilities of FECM are compared based on different fractional order-values. Furthermore, the reasonable fractional order-value interval is determined, which provides a reference value for the order-value selection in the fixed-order FECM. Test results show that for the two fractional order-values used in the 2-FRC equivalent model, when one is a fixed value, the other fluctuates in a moderate range in most cases. According to this principle, a fixed-order FECM is used to identify the battery parameters. On the basis of not losing the description accuracy of FECM, it reduces the computational complexity caused by fractional order operation.