A Stochastic Inference-Dual-Based Decomposition Algorithm for TSO-DSO-Retailer Coordination

Abstract

The flexibility services available from embedded resources, being attractive to both the network operators and retailers, pose a problem of co-ordination and market design at the local level. This research considers how a Flexibility Market Operator (FMO) at the local level, analogous to market operators at the wholesale level, can improve the real-time operation of the power systems and efficiently manage the interests of the TSO, DSO, and Retailers. Using generalized disjunctive programming, a stochastic bilevel representation of the task is reformulated as a stochastic mixed-logical linear program (MLLP) with indicator constraints. An Inference-Dual-Based Decomposition (IDBD) Algorithm is developed with sub-problem relaxation to reduce the iterations. Using expected Shapley values, a new payoff mechanism is introduced to allocate the cost of service activations in a fair way. Finally, the performance and benefits of the proposed method are assessed via a case study application.