Date of Award

May 2019

Degree Type


Degree Name

Master of Science



First Advisor

Lingfeng Wang

Committee Members

David Yu, Guangwu Xu


With the increasing consumption of energy, it is of high significance to improve energy efficiency and realize optimal operation of the multi-energy system. Among the many energy system modeling methods, the concept of “energy hub (EH)” is an emerging one. However, the previous EH models only included one or a few of constituting components.

The construction of an energy hub model that integrates energy storage systems, photovoltaic (PV) components, a combined cooling heating and power (CCHP) system and electric vehicles (EVs) is explained in this thesis. The inclusion of the CCHP system helps to meet the energy demand and improve the mismatch of heat-to-electric ratio between the energy hub and the load. Additionally, vehicle-to-grid (V2G) technology is applied in this EH; that is, EVs are regarded not only as load demands but also as power suppliers.

The energy hub optimization scheduling problem is formulated as a multi-period stochastic problem with the minimum total energy cost as the objective. Compared to 24-hour day-ahead scheduling, rolling horizon optimization is used in the EH scheduling and shows its superiority. In real-time rolling horizon scheduling, the optimization principle ensured that the result is optimized each moment, so it avoids energy waste caused by overbuying energy.

As part of electricity loads, EVs have certain influence on energy hub scheduling. However, due to the randomness of the driving patterns, it is still very difficult to perfectly predict the driving consumption and the charging availability of the EVs one day in advance. Chance constrained programming can hedge the risk of uncertainty for a big probability and drop the extreme case with a very low probability. By restricting the probability of chance constraints over a specific level, the influence of the uncertainty of electric vehicle charging behavior on energy hub scheduling can be reduced. Simulation results show that the energy hub optimization scheduling with chance constrained programming results in a less energy cost and it can make better use of time-varying PV energy as well as the peak-to-valley electricity price.