Authors: Zhang, R., Li, Z., Liu, P., Hawkes, A.D
Name of Journal: Sustainable Cities and Society
Abstract:Community energy systems globally are undergoing profound revolution towards sustainability, but face significant uncertainties from varying community conditions and differing preferences of decision-makers. While stochastic programming addresses parameter uncertainties effectively, growing attention has been directed towards “modelling to generate alternatives” (MGA), which provides diverse near-cost-optimal solutions to accommodate the varied needs of decision-makers beyond the limits of finite model structures. However, it is rarely recognized that these alternatives may differ significantly beyond economics, particularly in system resilience to renewable fluctuations, posing risks in achieving a sustainable and reliable community energy future through diversification. By introducing “modeling to generate resilience” (MGR), we propose a hierarchical algorithm and a quantile sampling to identify diverse and resilient alternatives, addressing both parameter uncertainty in renewables and structural uncertainty arising from model imperfections. With a campus community case, we find alternatives generated by tradition MGA may experience resilience degradation, while the modified algorithm ensures both diversity and resilience, reducing the average energy deficiency by 65%. Quantile sampling reveals four resilience characteristics within near-optimal space, navigating decision-makers in flexibly adjusting technology installations while ensuring system resilience. This offers practical insights for reliable energy infrastructure deployment under hybrid uncertainties incorporating diverse decision preferences and variable real-world conditions.
Full paper here