Optimizing Hierarchical Power Distribution of Multiple Local Energy Network Systems in Grid-Connected Mode

Abstract

IntroductionWith increasing environmental pollution and the reduction of fossil energy storage, decentralized renewable energy sources (DRES) have been considered to be effective alternatives for future energy requirements [1][2]. For adapting to the characteristics of DRES (e.g., geographical dispersion), local energy network systems (LENs) are employed to strengthen the collaborative relationship between DRES and the smart grid. LENs are composed of many small-scale but self-governing LENs whose basic members include renewable energy production units (e.g., wind turbines and photovoltaic generators), controllable storage units, and intelligent consumption units [3][4]. Different LENs are coupled Pol. J. Environ. Stud. Vol. 26, No. 5 (2017), 1981-2000 Original AbstractThis paper focuses on the hierarchical power distribution optimization of multiple local energy network (LEN) systems that are formed in three levels and can be operated in six typical modes. The decentralized optimal model for each LEN (the first level) and LENs (the second level) as well as the concentrated optimal model for the top level of the system are built, respectively. For each LEN, all the basic unities such as power generated by wind turbines and photovoltaic, and their upper nodes are considered. For LENs, the aggregated results (e.g., supply-demand requirements) from each LEN are dispatched. Furthermore, in the concentrated optimal control model (the third level), the ultimate supply-demand requirements of networked LENs together with other resources such as electric vehicles are considered. Due to the large number of control resources, the whole system is formulated as a large-scale global optimization (LSGO) problem. The self-adaptive differential evolution with neighborhood search method (SaNSDE) modified with the Lagrange multiplier method is used to solve the problem. The algorithm is firstly examined on 10 constrained benchmark functions, then it is applied to our problem. Experimental results show that the proposed model and algorithm are effective and efficient.Keywords: local energy network (LEN), energy internet (EI), decentralized renewable energy sources (DRES), adaptive neighborhood search technique, electric vehicles (EVs)