Clustering-Based Multi-objective Stochastic Optimization Approach in Scheduling Optimal Energy Dispatch of Multiple Energy Sources in the Presence of Uncertainty

The emerging trend of renewable and sustainable energy-based grid technologies are utilized to meet the increasing energy needs of highly energy intensive industries in particular. However, due to the intermittent nature of renewable energy sources as well as fluctuating energy prices, an effective characterization of the uncertain parameters become inevitable especially in the context of energy scheduling and planning. This work focuses on implementing the K-means clustering-based machine learning approach for characterizing the associated uncertain parameters in solving a multi-objective energy scheduling problem. In this study, in order to demonstrate the effectiveness of the aforementioned unsupervised machine learning approach, three energy sources are considered to be powering the load, among which few are prone to uncertainties. The hybrid energy sources are considered to deliver the requested energy for the flawless operation of the industrial facility. In essence, an appropriate quantity of scenarios of these stochastic parameters weighted by probabilities are generated using the K-means clustering approach which in turn gets utilized in generating the future energy dispatch schedule among the energy sources.