A two-stage stochastic MINLP model to design and operate a multi-energy microgrid by addressing carbon emission regulatory policies uncertainty

TL;DR

This paper develops a two-stage stochastic MINLP model to optimize multi-energy microgrid design and operation under uncertain carbon emission policies and renewable resource conditions, comparing deterministic and stochastic approaches.

Contribution

It introduces a novel stochastic MINLP model that incorporates policy and resource uncertainties for microgrid equipment selection and operation.

Findings

Stochastic model provides more robust equipment choices under policy uncertainty.

Optimal configurations differ significantly between deterministic and stochastic models.

Case studies demonstrate the model's effectiveness in diverse energy demand scenarios.

Abstract

This study suggests a novel two-stage Mixed-Integer Nonlinear Programming model considering uncertainty related to implementation of carbon dioxide emission regulatory policies, which are carbon trading and emission taxing and can change over the years, for the purpose of optimal equipment selection from candidate equipment to design, size and operate a multi-energy microgrid. The uncertain sources are air temperature, wind speed, solar radiation, carbon dioxide trading price or tax, and natural gas price. Candidate equipment are wind turbines, PV arrays, a biomass-fired generator, biomass combined cycles, combined heat and power generators, conventional generators, an electricity storage unit, integrated gasification combined cycles, a heat pump, and a power-to-synthetic natural gas (P2G) system. Three case studies are investigated. In the first case, the model selects the optimal equipment for meeting the electricity and heat demands only. In the second case, the optimal equipment selections are determined to couple with the P2G system to meet the electricity, heat, and natural gas demands. In the third case, the model selects the optimal equipment to run with sustainable energy generators: wind turbines and solar panels. The optimal selections are compared between deterministic and stochastic forms of the optimization models.