Micro Energy-Water-Hydrogen Nexus: Data-driven Real-time Optimal Operation

This paper extends a new concept of energy-water-hydrogen (EWH) nexus, which was recently developed as a solution for reducing carbon emissions from the generation side of power systems, to the distribution side. Under the concept of distribution-level EWH (micro EWH) nexus, renewable energy sources (RES) are utilized to meet the energy needs of a small community. To avoid the uncertainty caused by RESs, this paper aims to investigate the real-time optimal operation of the micro EWH nexus which is however a challenging optimization problem. First, such a large-scale mixed-integer nonlinear programming problem is relaxed into a mixed-integer convex program (MICP) by leveraging the effective convex-hull relaxation technique. Second, a fast data-driven solution method based on active constraint and integer variable prediction is presented, which can solve the MICP problem very fast since it utilizes historical optimization data to quickly predict binary variable values and a limited set of active constraints.