Optimizing electricity mix for CO2 emissions reduction: A robust input-output linear programming model

Abstract

Input-Output Linear Programming (IO-LP) model has been recently used to identify a cost-effective strategy for reduction in economy-wide CO2 emissions through a shift in the electricity generation mix. As an extension, this study further develops a robust IO-LP model to address the data uncertainties of technology cost and final demand. Compared to the deterministic IO-LP model which seeks to minimize the levelized cost of electricity (LCOE), the robust IO-LP model aims to maximize the tolerance of data uncertainty under a dynamic uncertainty setting. The modelling results in case study of China show that coal-fired and hydro generation technologies should be greatly developed from 2020 to 2050 in the Business-As-Usual (BAU) scenario with no emissions target set. In order to mitigate accumulated economy-wide CO2 emissions by 30% compared to the BAU emissions level, various types of clean generation technologies, i.e., gas-fired, hydro, nuclear, solar, wind, and biomass, should be introduced into the electricity mix. Along with the decrease in emissions target, the tolerance of data uncertainty will drop to a certain degree. Finally, we compared results of the robust IO-LP model with results of the stochastic and deterministic IO-LP models. The comparative analysis shows that the robust IO-LP model tends to select the generation technologies with smaller uncertainty in LCOE, and is able to improve the robustness of capacity planning solutions compared to the alternative models under data uncertainty.