A model-less control algorithm of DC microgrids based on feedback optimization

This work addresses the problem of optimal real-time control a DC microgrid without relying on its corresponding network model. The main goal such controller is to keep nodal voltages within regulatory limits while offering current sharing capability between different controllable generators powering microgrid. proposed model-less methodology based feedback optimization, which takes advantage available measurements update setpoints generation assets. variables are determined in an iterative manner by applying primal–dual saddle-point method, guarantees appropriate convergence features. paper details both centralized and distributed implementations compared through simulations. results evidence good dynamic performance steady-state operation as algorithm converges solution provided conventional model-based Optimal Power Flow.