Optimal Stochastic Self-Scheduling of a Water-Energy Virtual Power Plant Considering Data Clustering and Multiple Storage Systems

As the world moves towards the integration of different water and energy resources, as well as storage systems, it is necessary to develop the conventional structure of virtual power plants (VPPs). In this work, a mixed-integer nonlinear programming (MINLP) model is established for stochastic self-scheduling problem of a new VPP structure, namely water-energy VPP. The proposed VPP, which participates in the electricity market to maximize its daily profit, aggregates different energy and water resources including wind turbines, compressed air energy storage (CAES), gas boiler, electrical boiler, absorption chiller, ice storage, water storage, and water well to supply water demand, as well as different types of energy demands such as cooling, power, and heat demands. To cope with the uncertain behavior of wind power, a scenario-based approach is employed. First, with the use of Monte Carlo simulation in MATLAB, a number of scenarios are obtained, and afterwards, K-means technique, which is a fast and efficient data clustering method, is implemented for the scenario reduction. Additionally, the impact of the consideration of CAES system on the performance of the proposed water-energy is discussed. The numerical results indicate that with the deployment of CAES system into the structure of the water-energy VPP, the expected daily profit is increased up to 4.26 %.