Distributionally robust optimal dispatch of CCHP campus microgrids considering the time-delay of pipelines and the uncertainty of renewable energy

Considering the time-delay of energy transmission in cooling/heating pipelines and the uncertainty of renewable energy (RE), a distributionally robust optimal (DRO) dispatch model of combined cooling, heating and power campus microgrids (CCHP-CMG) is established. The method of characteristics is used to obtain the analytical algebraic solution of the partial differential equation (PDE) describing the energy transmission in pipelines and to add the solution to the optimal dispatch model. Based on the Jensen-Shannon divergence distance, an ambiguity set including the probability distribution (PD) information of the actual historical data is proposed to describe the uncertainty of RE outputs, and a min-max bi-level DRO dispatch model is established. An alternative iteration method is proposed to solve the bilevel model to obtain the solution that satisfies the constraints for the worst PD of uncertain variables. Additionally, a method for reconstructing the inner-layer problem is proposed to significantly reduce the computational complexity. Test results for an actual CCHP-CMG demonstrated the high computational accuracy and efficiency of the method of characteristics in solving the PDE compared to the finite difference method. The proposed reconstruction method is accurate and efficient, and the proposed DRO method can easily adjust the conservativeness of the obtained dispatch scheme. (c) 2021 Elsevier Ltd. All rights reserved.

Automated summary

A distributionally robust optimal dispatch model is established for campus microgrids, considering the time-delay of energy transmission and uncertainty of renewable energy. The method of characteristics is used to solve the partial differential equation, while a min-max bi-level DRO dispatch model is proposed based on an ambiguity set describing the uncertainty of RE outputs. The study also introduces an alternative iteration method for solving the bilevel model efficiently.