Resilient Decentralized Optimization Of Chance Constrained Electricity-Gas Systems Over Lossy Communication Networks

With the gradual growth of natural gas units, the coupling between power networks and natural gas networks has deepened, and the synergistic operation between them has become more and more important. At the same time, the integration of uncertain renewable energy brings challenges to the economic and safe operation of power and natural gas interconnected systems. In order to cope with the operational risks brought by uncertain wind sources to the interconnected system, this paper adopts the method of distributionally robust chance constraints, where an ambiguity set related to the moment information of a small amount of historical wind power forecast error data is attained in a data-driven way. Then the chance constrained problem is transformed into a formation that is easily solved. For different situations with or without a central coordinator of the two entities of electricity and gas, this work respectively gives the solution steps based on the relaxed alternating direction method of multipliers. In the case of no coordinator, this paper particularly presents the convergence performance when the exchange of information has a packet loss rate during iterative calculation processes. The simulation results show that compared with the method of traditional Gaussian chance constrained optimization and symmetrical distributionally robust chance constrained optimization, the moment-based distributionally robust chance constrained optimization can achieve a lowest probability of chance constraint violation at the expense of energy dispatch costs. In addition, compared with the traditional alternating direction method of multipliers, the relaxed alternating direction method of multipliers can achieve the nearly same energy dispatch costs with a smaller number of iterations. It is also found that the packet loss of the early iteration process may be the main factor that has a degradation impact on the convergence process rather than the nominal loss probability. (c) 2021 Published by Elsevier Ltd.