Optimal dispatch of clean resources in networked microgrids: A rank transactive energy framework

In this paper, a rank-based transactive energy framework is proposed to optimize the coordinated dispatch of clean local resources in networked microgrids. To enable effective coordination, the framework employs a ranking system to manage energy exchanges among microgrids. This approach minimizes load shedding while preserving the end-user privacy. The proposed framework leverages a novel learning-based probabilistic method to optimize the operation of networked microgrids using a distributionally robust, quantile-based chance-constrained approach. This eliminates the need for perfect knowledge of probability density functions by introducing a stochastic quantization technique for nonparametric probabilistic analyses. Quantiles are generated using a modified extreme learning machine autoencoder, trained through a two-phase process with historical data. To mitigate the computational burden associated with stochastic quantiles for numerous uncertainties, a step-by-step mixture process is employed, with the Lebesgue measure serving as the convergence criterion. The efficacy of the framework and the optimization approach is validated through several case studies and comprehensive comparisons. • A rank-based transactive framework coordinates energy dispatch in microgrids. • Learning-based probabilistic modeling handles dispatch uncertainty in a robust way. • The approach is distributionally robust, quantile-driven, and chance-constrained. • Stochastic quantization supports nonparametric, data-driven probabilistic analysis. • End-user privacy is preserved while minimizing load shedding across networks.