Uncertainty energy planning of net-zero energy communities with peer-to-peer energy trading and green vehicle storage considering climate changes by 2050 with machine learning methods

Abstract

This study aims to develop uncertainty energy planning of net-zero energy communities with innovative peer-to-peer energy trading management and advanced green vehicle storage considering climate changes by 2050 with transient system modelling, stochastic uncertainty sampling and data-driven machine learning predictions. The future-oriented peer-to-peer energy trading management strategy is proposed considering large renewable energy generation in determining time-of-use power periods to improve energy flexibility. The aleatory and epistemic uncertainties of key weather parameters determining renewables generation with climate changes are quantified and various machine learning regression models are used to predict dynamic solar photovoltaic and wind power generation by 2050 with weather uncertainty. The study results indicate that the ambient temperature increases with time at the uncertainty range of [0, 17.82%], the beam radiation and sky diffuse radiation decrease with time at [−24.86%, 0], and the ground reflected diffuse radiation and wind speed show an uncertain trend at [−0.72%, 0.72%]. The annual solar photovoltaic and wind power generation in 2050 is projected to be 11.69% and 0.61% lower than those of the typical meteorological year. The peer trading cost saving in 2050 considering climate changes is reduced by 2.25% and the annual equivalent carbon emissions are increased by 9.15% compared with the typical meteorological year scenario. The uncertainty design of net-zero energy communities integrated with hydrogen vehicle and battery vehicle storage considering climate changes provides significant guidance for advancing net zero carbon and realizing carbon neutrality for integrated building and transport sectors by 2050.