A Wasserstein based two-stage distributionally robust optimization model for optimal operation of CCHP micro-grid under uncertainties

Combined cooling, heating and power (CCHP) micro-grids are getting increasing attentions due to the realization of cleaner production and high energy efficiency. However, with the features of complex tri-generation structure and renewable power uncertainties, it is challenging to effectively optimize the operation of CCHP micro-grid. This paper proposed a novel Wasserstein based two-stage distributionally robust optimization (WTSDRO) model for the day-ahead optimal operation of CCHP micro-grid. The uncertainties of wind power (or other renewable energy sources with random power output) forecasting errors are modeled as an ambiguity set based on Wasserstein metric, which is assumed to contain all the possible probability distributions with a confidence level. In the first stage, CCHP micro-gird's operation cost is minimized according to the forecast information. In the second stage, for hedging against the perturbation of random wind power outputs, flexible resources are adjusted under the worst-case distribution within the ambiguity set. Multiple demand response programs (DRPs) are integrated to make electrical, thermal and cooling loads controllable. Finally, a reformulation approach is proposed based on strong duality theory, which equivalently transforms the WTSDRO model into a tractable MILP framework. Simulations implemented on a typical-structure CCHP micro-grid are delivered to show that our proposed model: (1) is data-driven and keeps both of the conservativeness and computational time at relatively low levels, (2) reaches effective operation results in terms of cost optimization, wind power accommodation and waste heat utilization etc. Moreover, operation cost and CO2 emission can be further saved by integrating multiple DRPs.