Deterministic and stochastic mixed-integer nonlinear programming for optimal design and operation of renewable-energy based microgrids to meet electricity, heat, and GAS demands

Abstract

This thesis consists of three parts. The first part suggests a novel deterministic one-layer and multi-period mixed-integer nonlinear programming (MINLP) model to select the equipment to install from a candidate pool for optimal design and scheduling of a microgrid to meet the electricity demand. The equipment selections are compared based on the implementation of carbon dioxide taxing and cap and trade system, utilization of seasonal and yearly average data of wind speed, air temperature, and global solar radiation, and grid modes. In the second part, a one-layer deterministic MINLP model is proposed for the equipment selection to be operated with a Powerto- Gas (PTG) system to meet the electricity and gas demands. The effects of carbon dioxide taxes and natural gas prices on equipment selections for a PTG-integrated microgrid are investigated. The third part proposes a two-stage stochastic MINLP model to meet the electricity, heat, and gas demands. Uncertainty sources are carbon dioxide regulation policies, weather data, and prices of natural gas and carbon dioxide. Three case studies are investigated. In the first case, the model selects the equipment for meeting the electricity and heat demands only. In the second case, the optimal selections are determined to couple with the PTG system. In the third case, the model chooses the equipment to run with sustainable energy generators. Finally, the differences in the optimal equipment selections between stochastic and deterministic models are compared.