Multi-year planning for the integration combining distributed energy system and electric vehicle in neighborhood based on data-driven model

As more residents choose electric vehicles (EVs), the strain on the neighborhood's energy supply will increase due to EVs’ dynamic charging requirements. Furthermore, the distributed energy system (DES) has demonstrated its benefit on local energy supply, considerably increasing its efficacy. This study concentrates on multi-year planning for the integration framework combining DES and EV charging supply in the neighborhood business center(NBC). Two types of EVs are considered, and their models are developed separately based on different stochastic behaviors. A novel data-driven method using real weather data to generate charging scenarios is proposed, which can provide quantitative guidance for planning. Latin hypercube sampling(LHS) is applied as the technical method for scenario generating. Machine learning(ML) clustering method with elbow criterion is used to conduct scenarios reduction. The multi-year planning optimization of the integration is accomplished based on two-stage stochastic programming. The multi-year planning project is employed in a NBC in Beijing. The planning result of integration cases with different external power limitations is assessed comprehensively. Other planning methods are introduced and compared. It shows that the integration case with moderate limitation can achieve a reduction of 67.8% and 31.6% respectively in economic cost and carbon emission. Furthermore, the EV satisfaction rate in varied scenarios is evaluated. What is more, the integration cases are furtherly assessed based on varied EV growth expectations. In general, the results have verified the effectiveness of the multi-year planning in neighborhood.