PV-Microgrid Operational Cost Minimization by Neural Forecasting and Heuristic Optimization

Advantages provided by the renewable based distributed generation coupled with federal and state deregulation policies and incentives, are pushing the future energy markets to invest more into renewable systems. Microgrid is a well-known concept for integrating the distributed resources into the current power distribution network. With advanced power electronics and controls, the technology already exists to develop microgrid with renewable energy sources. But, still the biggest concern in the mind of general consumers lies in the form of cost of energy of these renewable systems. Installation of the renewable energy systems requires high initial costs. In addition, the analysis and deployment of renewable energy system is challenging because of large number of design options, uncertainty in future fuel price, intermittent and seasonal nature of renewable power generation. Irrespective of interconnection technology, the optimization of those power electronics based system is extremely important to minimize the operation cost. In this paper, a distributed intelligent energy management system (DIEMS) is implemented to optimize operating costs of a representative photovoltaics (PV) based microgrid. Depending on the fuzzy ARTMAP neural network based forecast of PV generation, an optimization scheme is developed utilizing linear programming along with heuristics. Case studies are done for a household to demonstrate that the proposed DIEMS not only reduces the cost of operation, but also helps to improve overall system operation such as lower maintenance for storage and improved battery life.