RHO–LSTM-based optimal scheduling at the motorcycle battery swapping station under battery heterogeneity

Authors:
Open Access
Publisher website
Google Scholar
Abstract
This research proposes a mechanism that enables the battery swapping station (BSS) to provide battery swap services for multiple types of batteries, termed battery heterogeneity, utilized in electric motorcycles. The number of batteries for each type is established. The battery charging cost is calculated in real time, and the station's profit is maximized by optimizing battery swap scheduling. The issues are modeled as a mixed-integer non-linear problem (MINLP), then linearized as a mixed-integer linear problem (MILP), using the grid electricity price from the real-time pricing mechanism to calculate the battery's charging/discharging cost. Swap scheduling is optimized using the rolling horizon optimization (RHO) approach, which takes into account a variety of constraints. These constraints include battery type, battery SoC, arrival time of the electric motorcycle, grid electricity pricing at time t, and battery power utilization. The long-short term memory (LSTM) predicts the electric motorcycles' arrival time at t+1 based on prior data. The results show that optimization scheduling generates a higher overall profit per day than unscheduled operation. Profit by the RHO-LSTM method is 23.77 % greater than by the RHO-Polynomial method and 0.26 % greater than by unscheduled operation. Furthermore, the number of batteries provided by the RHO-LSTM method is 40 % greater than by the RHO-polynomial method.

This publication has 29 references indexed in Scilit: