An Adaptive Decentralized Economic Dispatch Method For Virtual Power Plant

This paper introduces a decentralized economic dispatch method and an architecture suitable for the virtual power plant (VPP) aggregating massive distributed energy resources (DERs). The convergence condition is given for quadratic cost functions, and is extended to the case of general increasing function of incremental cost (IC). Further analysis shows that the step of this method is adaptive, which is generated from the bottom up according to the responsiveness of each DER unit (DERU). Combined with the decentralized architecture based on message queue (MQ), the algorithm design considers the hosting mechanism of the coordinator failure, which not only improves the efficiency of calculation and communication without losing privacy-protection, but also makes it more fault-tolerant. The correctness and effectiveness of the method are verified in the case studies. The iterative process can respond and converge quickly when DER units reach capacity limits or devices fail/join. Due to the adaptability of the step, the method has strong robustness to the quantity and parameters randomness of underlying units. Therefore, it can be applied to the VPP with a massive number of DERs in order to get consensus solution by rapid economic dispatch.